Front Cover
 Title Page
 Table of Contents
 Contributor's biographies
 Abbreviations and acronyms
 Part 1: FSR- Understanding farmers...
 Part 2: The applications of farming...
 Part 3: Institutional commitment...
 Part 4: FSR- The professional...
 Part 5: Cutting edge methods, abiding...
 Back Cover

Group Title: history of farming systems research
Title: A history of farming systems research
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00085336/00001
 Material Information
Title: A history of farming systems research
Physical Description: xvi, 432 p. : ill. ; 25 cm.
Language: English
Creator: Collinson, M. P ( Michael P )
Publisher: Food and Agriculture Organization of the United Nations
Place of Publication: Rome Italy
New York
Publication Date: c2000
Subject: Agricultural systems -- Research -- History   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Bibliography: Includes bibliographical references and index.
Statement of Responsibility: edited by M. Collinson.
 Record Information
Bibliographic ID: UF00085336
Volume ID: VID00001
Source Institution: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 42021543
lccn - 99042827
isbn - 0851994059 (alk. paper)

Table of Contents
    Front Cover
        Front Cover
    Title Page
        Page i
        Page ii
    Table of Contents
        Page iii
        Page iv
        Page v
    Contributor's biographies
        Page vi
        Page vii
        Page viii
        Page ix
    Abbreviations and acronyms
        Page x
        Page xi
        Page xii
        Page xiii
        Page xiv
        Page xv
        Page xvi
        Page 1
        Page 2
        Page 3
        Page 4
    Part 1: FSR- Understanding farmers and their farming
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        FSR: Origins and perspectives
            Page 13
            Page 14
            Page 15
            Page 16
            Page 17
            Page 18
            Page 19
            Page 20
            Page 21
            Page 22
            Page 23
            Page 24
            Page 25
            Page 26
            Page 27
            Page 28
            Page 29
            Page 30
            Page 31
            Page 32
            Page 33
            Page 34
            Page 35
            Page 36
            Page 37
            Page 38
            Page 39
            Page 40
        FSR: Understanding farming systems
            Page 41
            Page 42
            Page 43
            Page 44
            Page 45
            Page 46
            Page 47
            Page 48
            Page 49
            Page 50
            Page 51
            Page 52
            Page 53
            Page 54
            Page 55
            Page 56
            Page 57
            Page 58
            Page 59
            Page 60
            Page 61
            Page 62
            Page 63
            Page 64
            Page 65
            Page 66
            Page 67
            Page 68
            Page 69
            Page 70
            Page 71
            Page 72
            Page 73
            Page 74
            Page 75
            Page 76
            Page 77
            Page 78
            Page 79
            Page 80
            Page 81
            Page 82
    Part 2: The applications of farming systems research
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        FSR in technology choice and development
            Page 95
            Page 96
            Page 97
            Page 98
            Page 99
            Page 100
            Page 101
            Page 102
            Page 103
            Page 104
            Page 105
            Page 106
            Page 107
            Page 108
            Page 109
            Page 110
            Page 111
            Page 112
            Page 113
            Page 114
            Page 115
            Page 116
            Page 117
            Page 118
            Page 119
            Page 120
            Page 121
            Page 122
            Page 123
            Page 124
            Page 125
            Page 126
            Page 127
            Page 128
            Page 129
            Page 130
            Page 131
            Page 132
            Page 133
            Page 134
            Page 135
            Page 136
            Page 137
            Page 138
        FSR in extension and policy formation
            Page 139
            Page 140
            Page 141
            Page 142
            Page 143
            Page 144
            Page 145
            Page 146
            Page 147
            Page 148
            Page 149
            Page 150
            Page 151
            Page 152
            Page 153
            Page 154
            Page 155
            Page 156
            Page 157
            Page 158
            Page 159
            Page 160
    Part 3: Institutional commitment to farming systems research
        Page 161
        Page 162
        Page 163
        Page 164
        Page 165
        Page 166
        Page 167
        Page 168
        FSR: Some institutional experiences in national agricultural research
            Page 169
            Page 170
            Page 171
            Page 172
            Page 173
            Page 174
            Page 175
            Page 176
            Page 177
            Page 178
            Page 179
            Page 180
            Page 181
            Page 182
            Page 183
            Page 184
            Page 185
            Page 186
            Page 187
            Page 188
            Page 189
            Page 190
            Page 191
            Page 192
            Page 193
            Page 194
            Page 195
            Page 196
            Page 197
            Page 198
            Page 199
            Page 200
        Some dimensions of the organization of FSR
            Page 201
            Page 202
            Page 203
            Page 204
            Page 205
            Page 206
            Page 207
            Page 208
            Page 209
            Page 210
            Page 211
            Page 212
            Page 213
            Page 214
            Page 215
            Page 216
            Page 217
            Page 218
            Page 219
            Page 220
            Page 221
            Page 222
            Page 223
            Page 224
        Training for FSR
            Page 225
            Page 226
            Page 227
            Page 228
            Page 229
            Page 230
            Page 231
            Page 232
            Page 233
            Page 234
            Page 235
            Page 236
            Page 237
            Page 238
            Page 239
            Page 240
            Page 241
            Page 242
            Page 243
            Page 244
            Page 245
            Page 246
    Part 4: FSR- The professional dimension
        Page 247
        Page 248
        Page 249
        Page 250
        The regional and international associations
            Page 251
            Page 252
            Page 253
            Page 254
            Page 255
            Page 256
            Page 257
            Page 258
            Page 259
            Page 260
            Page 261
            Page 262
            Page 263
            Page 264
            Page 265
            Page 266
            Page 267
            Page 268
            Page 269
            Page 270
            Page 271
            Page 272
            Page 273
            Page 274
            Page 275
            Page 276
            Page 277
            Page 278
            Page 279
            Page 280
            Page 281
            Page 282
            Page 283
            Page 284
            Page 285
            Page 286
            Page 287
            Page 288
            Page 289
            Page 290
            Page 291
            Page 292
        FSR and the professional disciplines
            Page 293
            Page 294
            Page 295
            Page 296
            Page 297
            Page 298
            Page 299
            Page 300
            Page 301
            Page 302
            Page 303
            Page 304
            Page 305
            Page 306
            Page 307
            Page 308
            Page 309
            Page 310
            Page 311
            Page 312
            Page 313
            Page 314
            Page 315
            Page 316
            Page 317
            Page 318
    Part 5: Cutting edge methods, abiding issues, and the future for FSR
        Page 319
        Page 320
        Page 321
        Page 322
        At the cutting edge
            Page 323
            Page 324
            Page 325
            Page 326
            Page 327
            Page 328
            Page 329
            Page 330
            Page 331
            Page 332
            Page 333
            Page 334
            Page 335
            Page 336
            Page 337
            Page 338
            Page 339
            Page 340
            Page 341
            Page 342
            Page 343
            Page 344
            Page 345
            Page 346
            Page 347
            Page 348
            Page 349
            Page 350
            Page 351
            Page 352
            Page 353
            Page 354
            Page 355
            Page 356
            Page 357
            Page 358
            Page 359
            Page 360
            Page 361
            Page 362
            Page 363
            Page 364
            Page 365
            Page 366
            Page 367
            Page 368
            Page 369
            Page 370
            Page 371
            Page 372
            Page 373
            Page 374
            Page 375
            Page 376
            Page 377
            Page 378
            Page 379
            Page 380
            Page 381
            Page 382
            Page 383
            Page 384
            Page 385
            Page 386
            Page 387
            Page 388
            Page 389
            Page 390
        The future of farming systems research
            Page 391
            Page 392
            Page 393
            Page 394
            Page 395
            Page 396
            Page 397
            Page 398
            Page 399
            Page 400
            Page 401
            Page 402
            Page 403
            Page 404
            Page 405
            Page 406
            Page 407
            Page 408
            Page 409
            Page 410
            Page 411
            Page 412
            Page 413
            Page 414
            Page 415
            Page 416
            Page 417
            Page 418
            Page 419
        Page 420
        Page 421
        Page 422
        Page 423
        Page 424
        Page 425
        Page 426
        Page 427
        Page 428
        Page 429
        Page 430
        Page 431
        Page 432
    Back Cover
        Page 433
Full Text

history of Farming

All ,-liuu --ur--- uw



Edited by
M. Collinson

Published by
Food and Agriculture Organization of the United Nations
CABI Publishing

M. Collinson and FAO 2000.
A catalogue record for this book is available from the British Library, London, UK.
Library of Congress Cataloging-in-Publication Data
A history of farming systems research / edited by M.P. Collinson.
p. cm.
Includes bibliographical references (p. ).
ISBN 0-85199-405-9 (alk. paper)
1. Agricultural systems--Research History. I. Collinson, M.P.
(Michael P.)
S494.5.S95H57 1999
630'.72--dc21 99-42827
Published jointly by:
CABI Publishing, a division of CAB International
CABI Publishing, CAB International, Wallingford, Oxon OX10 8DE, UK
Tel: +44(0)1491 832111, Fax: +44(0)1491 833508, Email: cabi@cabi.org
CABI Publishing, 10 E 40th Street, Suite 3203, New York, NY 10016, USA
Tel: +1 212 481 7018, Fax: +1 212 686 7993, Email: cabi-nao@cabi.org
Food and Agriculture Organization of the United Nations (FAO),
Viale delleTerme di Caracalla, 00100 Rome, Italy
Tel: +39 06 57051, Fax: +39 06 57053152, Email: webmaster@fao.org
Website: http://www.fao.org
ISBN 0 85199 405 9 (CABI) hard cover edition
ISBN 92 5 104311 6 (FAO) paperback edition
The designations employed and the presentation of material in this publication
do not imply the expression of any opinion whatsoever on the part of the Food
and Agriculture Organization of the United Nations concerning the legal status
of any country, territory, city or area or of its authorities, or concerning the
delimitation of its frontiers or boundaries.
The designations 'developed' and 'developing' economies are intended for
statistical convenience and do not necessarily express a judgement about the
stage reached by a particular country, territory or area in the development process.
The views expressed herein are those of the authors and do not necessarily
represent those of the Food and Agriculture Organization of the United Nations.
All rights reserved. No part of this publication may be reproduced in any form or
by any means, electronically, mechanically, by photocopying, recording or other-
wise, without the prior permission of FAO.
Typeset by Columns Design Ltd, Reading.
Printed and bound in the UK by Biddies Ltd, Guildford and King's Lynn


Contributors' Biographies vi
Abbreviations and Acronyms x

Foreword xii
Janice Jiggens

Chapter 1 Introduction 1
Mike Collinson

Editorial Introduction 5
Mike Collinson

Chapter 2 FSR: Origins and Perspectives 13
2.1 My initiation into FSR in Latin America 13
German Escobar
2.2 A personal history in FSR 18
Peter Hildebrand
2.3 The evolution of FSR-E in Asia through the mid 1970s: 22
a view from IRRI
Richard Harwood
2.4 FSR: a personal evolution 30
David Norman
2.5 My FSR origins 34
Mike Collinson

Chapter 3 FSR Understanding Farming Systems 41
3.1 FSR's expanding conceptual framework 41
Robert Hart
3.2 Evolving typologies for agricultural R & D 51
Mike Collinson
3.3 The development of diagnostic methods in FSR 59
John Farrington


3.4 Gender analysis: making women visible and improving social analysis 67
Hilary Sims Feldstein
3.5 Relating problems and causes in FSR planning 76
Robert Tripp

Editorial Introduction 83
Mike Collinson

Chapter 4 FSR in Technology Choice and Development 95
4.1 The application of FSR to technology development 95
Ann Stroud and Roger Kirkby
4.2 Experiences in applying FSR in semi-arid Kenya 130
A.J. Sutherland and J.N. Kang'ara

Chapter 5 FSR in Extension and Policy Formulation 139
5.1 Farming systems extension in the USA 139
Cornelia Butler Flora and Charles Francis
5.2 The evolution of the advisory services in Chile and the role of FSR 145
Julio A. Berdegu6
5.3 A farming systems contribution to agricultural policy analysis 152
John Dixon

Editorial Introduction 161
Mike Collinson

Chapter 6 FSR: Some Institutional Experiences in National
Agricultural Research 169
6.1 The Systems Research Department at INRA 169
J. Bonnemaire et al.
6.2 Senegal's experimental units 178
Michel Benoit-Cattin
6.3 Twenty years of systems research in southern Mali the Sikasso 184
FSR experience
Willem Stoop et al.
6.4 The institutionalization of FSR in east and southern Africa: an overview 191
Stuart A. Kean and M. Creasy Ndiyoi

Chapter 7 Some Dimensions of the Organization of FSR 201
7.1 Institutionalizing FSR in Tanzania: a case study 201
Ann Stroud
7.2 Institutionalizing FSR in Zambia: a stakeholder perspective 208
Stuart A. Kean and M. Creasy Ndiyoi
7.3 Costs of on-farm research: a comparison of experiences 215
in six countries
Elon H. Gilbert

Chapter 8 Training for FSR 225
8.1 The history of FSR training in east, central and 225
southern Africa
Ponniah Anandajayasekeram


8.2 Orienting research to agricultural development: 234
the ICRA training programme
Richard Hawkins
8.3 A note: the story behind the 'guidelines' 242
William W. Shaner

Editorial Introduction 247
Mike Collinson

Chapter 9 The Regional and International Associations 251
9.1 Ten years in the making: the Association for Farming 251
Systems Research and Extension
Hal MacArthur
9.2 Farming Systems Research and Extension in Latin America 261
Julio A. Berdegud
9.3 An overview of FSR-E and FSR-E networks in Africa 277
James Olukosi
9.4 The Asian Farming Systems Association 286
Nimal Ranweera

Chapter 10 FSR and the Professional Disciplines 293
10.1 Farm management and the farming systems approach 293
David Norman
10.2 Anthropology, sociology and FSR 299
Constance M. McCorkle
10.3 Agronomy and FSR a reluctant marriage? 312
Peter Hildebrand and Dennis Keeney

Editorial Introduction 319
Mike Collinson

Chapter 11 At the Cutting Edge 323
11.1 Holism and FSR 323
Toon van Eijk
11.2 The GIS and remote sensing contribution to the elaboration of system 334
hierarchies in FSR
Evaristo Miranda
11.3 FSR from a modelling perspective: experiences in Latin America 342
Roberto A. Quiroz et al.
11.4 Moving participatory plant breeding forward: the next steps 354
Louise Sperling and Jacqueline A. Ashby
11.5 Agroecosystems analysis: a systems application with a future? 363
Clive Lightfoot
11.6 Water quality, agricultural practices and changes in farming and 382
agrarian systems
J.P. Deffontaines et al.
Chapter 12 The Future of Farming Systems Research 391
Mike Collinson and Clive Lightfoot


Contributors' Biographies


Mike Collinson. The first farm economist
appointed into the British colonial agricultural
research services (1960), Mike worked with
resource-poor African farmers for 25 years. From
1975, as regional economist with CIMMYT, he
helped build capacity for on-farm research with
systems perspective in the agricultural research
institutions of eastern and southern Africa. He
retired to the UK in 1996 after 9 years as social
science adviser in the secretariat of the
Consultative Group for International Agricultural
Research (CGIAR), in Washington, DC.
Email: mikecollinsonl@compuserve.com

Ponniah Anandajayasekeram. Ananda, as
he is universally known in farming systems
research circles, has directed FSR training pro-
grammes and advised on methods and institu-
tional capacity building in Africa since 1982.
He was a Past President of the Southern
African Association for FSR-E and the immedi-
ate Past President of the global Association for
Farming Systems Research/Extension (AFSR/E).
Email: pananda@farmesa.co.zw
Michel Benoit-Cattin. Michel is in charge of
the international scientific exchanges in CIRAD,
in Montpellier, France. An agronomist with a
State Doctorate in economics he has worked for
30 years on agricultural development problems
at the farm, regional and national levels in
France, Algeria, Ivory Coast and Senegal. He
now supervises theses in these topics by young

researchers from all parts of the world.
Email: benoitca@cirad.fr
Julio A. Berdegu6. Based in Chile, Julio has
been the Coordinator of RIMISP, the
International Farming Systems Network, since
its foundation in 1986. Between 1992 and
1995 he was the Technical Director of Chile's
Agricultural Development Institute (INDAP).
Julio is the co-editor of four books on different
aspect of FSR-E, and is the owner and manager
of a seed and vegetable producing farm.
Email: berdegue@reuna.cl
Joseph Bonnemaire, Jacques Brossier
and B. Hubert. All three are scientists with
Systems Agraire et Development (SAD), the
unique agrarian systems division of INRA in
France. Jacques Brossier was a co-organizer of
the 1994 Montpellier AFSRE Symposium and
is Chair of the INRA Regional Research
Centre of Dijon. B. Hubert is a veterinarian
and a department chief in INRA. Joseph
Bonnemaire is an animal husbandry special-
ist, a professor in ENESAD and a 'charge de
mission' with INRA.
Email: brossier@dijon.inra.fr
Cornelia Butler-Flora and Charles Francis.
Cornelia is Director of the North Central Regional
Center for Rural Development, which conducts
research and extension programmes in rural devel-
opment in the 12 midwestern American states,
and is Professor of Sociology at Iowa State
University. Charles is Agronomist and Director of
the Center for Sustainable Agricultural Systems,
University of Nebraska, Lincoln. Cornelia and
Charles have pioneered the application of a systems
approach in extension in the American Midwest.

Contributors' Biographies

Emails: cflora@iastate.edu charles.francis@ipf.
J.P. Deffontaines, Jacques Brossier, M.
Barbier, M. Benoit, E. Chia, J.L. Fiorelli, M.
Gafsi, F. Gras, H. Lemery and M. Roux. The
authors are members of the SAD research team
in INRA responsible for the Vittel study. Brossier
and Deffontaines had scientific responsibility for
the study. J.P. Deffontaines is the father of 'geo-
agronomy' concerned with spatial connections
of farming systems and played a leading role in
the establishment of the SAD department in
INRA in 1979. Lemery is a sociologist specializ-
ing in farmers' behaviour and social networking
with a focus on the transformation of profes-
sional practices relating to agriculture. He is a
professor in ENESAD, Dijon. Roux is an animal
specialist, and a professor in ENESAD, Dijon.
Barbier, Chia and Gafsi are economists, Benoit
and Fiorelli are agronomists. Gras was researcher
in biological pedology at CNRS, the University of
Nancy and ORSTOM. He is now retired.
Email: brossier@dijon.inra.fr
John Dixon. John is Programme Coordinator,
Farm-Level Agricultural Research Methods for
East and Southern Africa, Harare, Zimbabwe.
He has worked in farming systems research and
development for more than 25 years, half of
which has been spent in Ethiopia, Iran, Nepal,
Thailand and Zimbabwe. He was also based in
Rome for several years implementing FAO's
farming systems programme.
Email: john.dixon@farmesa.co.zw
German Escobar. German is the Research
Director of the International Network on
Farming Systems Research Methodology (RIM-
ISP), based in Chile. A Colombian researcher,
German trained as an Agricultural Economist
and FSR practitioner and has worked in several
countries as researcher, project designer and
advisor to national research and rural develop-
ment institutions.
Email: ge@rimisp2.cl
John Farrington. John heads the Rural Policy
and Environment Group at the Overseas
Development Institute, London. UK, and is Visiting
Professor at the Agricultural Extension and Rural
Development Department of Reading University.
Email: j.farrington@odi.org.uk
Elon Gilbert. A Visiting Research Fellow with
the Overseas Development Institute, Elon is an
agricultural economist with more than 20
years experience in Sub-Saharan Africa and,
more recently, in South Asia. He wrote, with
others, the original American Journal of
Agricultural Economics article on FSR in 1980,

and was a member of the ISNAR team examin-
ing national programme experience with FSR in
the mid 1980s.
Email: elonsm@aol.com
Robert Hart. While working at CATIE, Bob, an
ecologist, proposed the multilevel systems hier-
archy now accepted as a conceptual framework
for FSR. He is currently the Director of the
USAID-funded global Sustainable Agriculture
and Natural Resource Management (SANREM)
Collaborative Research Support Program coor-
dinated by the University of Georgia.
Email: rdhart@arches.uga.edu
Richard Harwood. The founder and Head of
IRRI's Asian Cropping Systems Program from
1967 to 1972, Richard developed the Asian
Cropping Systems Network. At Rodale, he
directed the Research Center and its integrated
systems studies until 1985. As Director of
Asian Programes for Winrock International
from 1985 to 1990, he supported farming sys-
tems projects in Indonesia, Nepal and the
Philippines. Currently the C.S. Mott Chair of
Sustainable Agriculture at Michigan State
University, he is also a member of the CGIAR
Technical Advisory Committee.
Email: rharwood@pilot.msu.edu
Richard Hawkins. Following his PhD studies
on maize growth and yield in the Kenya high-
lands, Richard spent 10 years on FSR projects
in Central America (CATIE), Nepal and Java.
Richard has spent the last 10 years putting this
experience to good use, first as Coordinator of
the ICRA Anglophone training programme in
research for development, and more recently
with the Postgraduate College in Mexico.
Email: rhawkins@infosel.net.mx
Peter Hildebrand and Dennis Keeney. Peter

is Professor, Food and Resource Economics
Department, Institute of Food and Agricultural
Sciences, University of Florida, Gainesville. He
has worked in farming systems research for
more than 25 years, including 13 years in
Colombia, El Salvador and Guatemala. He has
coordinated the University of Florida Farming
Systems Research-Extension Program for 15
years and was the founding President of the
global Association for Farming Systems
Research-Extension. Dennis Keeny, a past
President of the American Agronomy Society, is
currently Director of the Leopold Center for
Sustainable Agriculture, Iowa State University.
Email: hildebrand@fred.ifas.ufl.edu
Janice Jiggens. A past President of the
International Association for Farming Systems

viii Contributors' Biographies

Research and Extension, Janice is currently
Professor of Human Ecology at the Swedish
Agricultural University. She has spent most of
her interdisciplinary career working on agricul-
ture and rural development, mainly in Africa
and South Asia. She has a particular interest in
combining participatory and 'high tech'
approaches and tools, and in ensuring that
gender issues, and women professionals, are
fully present in the agricultural sciences.
Email: janice.jiggens@lbutv.slu.se
Stuart Kean and M. Creasy Ndiyoi. Stuart is
a social scientist with interests in agriculture,
rural development and community-based nat-
ural resource management. He worked for 11
years in Zambia and was the first National
Coordinator of the Adaptive Research Planning
Team (ARPT). He is currently coordinating the
Northern Namibia Environmental Project fea-
turing multi-agency community-based natural
resource management initiatives.
Email: skean@iafrica.com.na
Ndiyoi is the Chief Agricultural Research
Officer, Farming Systems and Social Sciences,
Ministry of Agriculture Food and Fisheries,
Zambia. He has worked in the field of farming
systems and small farmer development for the
last 17 years. He is the current President of the
Southern African Association for Farming
Systems Research and Extension (SAAFSRE).
Email: mndiyoi@pop3.zamnet.zm
Clive Lightfoot. Clive, an agronomist, worked
in on-farm research for many years in Africa
and in Asia. He has put much of the flesh on
the bones of agroecosystems research and is
Chairman of the International Support Group
(ISG) for stakeholder reconciliation.
Email: clive.lightfoot@agropolis.fr
Hal MacArthur. A past President of the
International Association for Farming Systems
Research, Hal has worked extensively in FSR
programmes in both Africa and Asia.
Email: hmcarthur@hawaii.edu
Constance McCorkle. A sociologist with wide
experience in Africa, Asia and Latin America,
Constance is a world authority on the study and
application of ethnoveterinary medicine.
Email: constancemckorkle@msn.com
Evaristo Eduardo de Miranda. Born in
Brazil, Evaristo has had over 100 works pub-
lished at home and abroad. He is a Professor at
SAo Paulo State University (USP), researcher at
the EMBRAPA Remote Sensing Monitoring
Center (www.nma.embrapa.br) and also heads
the NGO ECOFORCE Research and Development

(www.ecof.org.br). He has applied remote sens-
ing and geographic information systems to the
assessment and monitoring of development and
environmental protection programmes.
Email: mir@nma.embrapa.br
David Norman. David is Professor.
Department of Agricultural Economics, Kansas
State University. Since the mid 1960s he has
spent a total of 20 years working in national
agricultural research systems in Africa (Nigeria
and Botswana) and continues to take on short-
term assignments in Africa, Asia and the South
Pacific. He was among those who initiated the
annual farming systems research and extension
symposia at Kansas State University, which
later evolved into the global Association for
Farming Systems Research-Extension (now the
International Farming Systems Association).
He is also a past President of the Association.
Email: dnorman@agecon.ksu.edu
James Olukosi. James is a professor of
Agricultural Economics with the Institute of
Agricultural Research at Ahmadu Bello
University in Nigeria, and has worked widely in
FSR in West Africa. He is the Coordinator of the
West African Farming System Research
Network (WAFSRN).
Email: ICRISAT-W-NIGERIA@cgiar.org
Roberto A. Quiroz, Carlos Leon-Velarde
and Walter Bowen. Roberto is the Head of the
Production Systems and Natural Resource
Management Department at the International
Potato Center (CIP). He received his MS and
PhD degrees from North Carolina State
University. His interests are in FSR and in the
integration of GIS and remote sensing with sim-
ulation models in land use studies. Carlos is a
systems scientist at the CIP and the
International Livestock Research Institute
(ILRI). He holds a PhD from the University of
Guelph in animal breeding and genetics. His
research has focused on the development and
application of simulation models in mixed
crop-livestock systems, and natural resources.
Walter is a systems scientist, specializing in soil
fertility, with the International Fertiliser
Development Center (IFDC), outposted to the
CIP. He earned his MS and PhD degrees from
Cornell University. His research activities have
centred on the development and application of
crop growth models.
Email: R.Quiroz@cgiar.org
Nimal Ranweera. Nimal is a past president of
both the Asian Association for Farming Systems
Research and the International Association for
Farming Systems Research-Extension. He is a

Contributors' Biographies

senior administrator in the Ministry of
Agriculture, Sri Lanka. He hosted and orga-
nized the 1996 international FSRE symposium.
Email: minagr@slt.lk
Bill W. Shaner. Bill is Professor Emeritus,
Colorado State University. He was team leader
and senior author of the book, Farming Systems
Research and Development: Guidelines for Developing
Countries. He occasionally consults overseas on
projects related to economic development.
Email: interdev@lamar.colostate.edu
Hilary Sims Feldstein. Hilary is the Training
Specialist at the International Center for
Research on Women, Washington, DC. For 9
years, she was the Program Leader, Gender
Analysis, for the Gender Program at the CGIAR.
In recent years she has worked with the Kenya
Agricultural Research Institute and with
Uganda's National Environmental Management
Authority to integrate gender analysis into their
organizations' technology development and
natural resource management activities.
Email: HilaryFeldstein@dai.com
Louise Sperling and Jacqueline Ashby.
Louise Sperling, an anthropologist, is a senior
scientist at CIAT and facilitates the Plant
Breeding Group of the CGIAR Systemwide
Program on Participatory Research for Gender
Analysis and Technical Innovation. She has
worked for 20 years in Africa and Asia on inno-
vative breeding and seed system strategies to
benefit small farmers. Jacqueline Ashby is a
PhD in development sociology who has pub-
lished extensively on participatory research and
social ecology. She is currently Director of
Research for Natural Resource Management at
the International Center for Tropical
Agriculture (CIAT) where she previously devel-
oped approaches to applied participatory
research which are now taught and practised in
several countries across the world.
Email: L.Sperling@cgiar.org J.Ashby@cgiar.org
Willem Stoop, Omar Niangado, Demba K6b6
and Toon Defoer. Willem is an agronomist and
soil scientist. Since 1992 he has been external
advisor to L'Institut d'Economie Rurale in
Bamako, Mali, and its respective ESPGRN teams,
including Sikasso. He was formerly with the
Royal Tropical Institute in Amsterdam and is
now an independent consultant. Omar Niangado
is a plant breeder with particular expertise in mil-
let. He was Director General of L'Institut
d'Economie Rurale in Bamako, Mali, from 1993
to 1998. Demba K6b6 is agricultural economist
at L'Institut d'Economie Rural in Bamako and
was formerly Head of the ESPGRN team in

Sikasso. Toon Defoer, Senior Scientist at the Royal
Tropical Institute (KIT) Amsterdam, specializes
in famer participatory methodology development
for natural resources management. He was the
technical advisor to the ESPGRN team in Sikasso
from 1993 to 1996. Email: nadamo@wxs.nl
Ann Stroud and Roger Kirkby. Ann, a
Californian, has resided in East Africa since
1982. She has worked for a number of organi-
zations specializing in weed management, pesti-
cide issues, farming systems research and more
recently natural resource management. Her
experience in systems agronomy is derived from
field research and extensive advising and capac-
ity building with NARS in the East and south-
ern African region. She is currently the
coordinator of the African Highlands Initiative,
an ecoregional programme at ICRAF. Roger is
an agronomist who has carried out on-farm
experiments in eastern Africa and Latin
America since 1969. His PhD from Cornell
University included looking at ways to integrate
experimentation by farmers and the formal sec-
tor; other degrees are from the Universities of
Wales and Cambridge. He is currently CIAT's
Coordinator for Africa.
Email: A.Stroud@cgiar.org R.Kirkby@cgiar.org
Alistair Sutherland and John N. Kang'ara.
John Kang'ara is a researcher in livestock pro-
duction and nutrition with the Kenya
Agricultural Research Institute, based at Embu
Research Centre. He was the livestock
researcher on the DAREP project from 1993 to
1997 and the project coordinator from 1995.
Alistair Sutherland is a social anthropologist
with the Natural Resources Institute of the
University of Greenwich, UK. He was the
anthropologist and technical advisor on the
DAREP project from 1993 to 1997.
Email: alistair.sutherland@nri.org

Robert Tripp. Robert is an anthropologist and
currently a research fellow at the Overseas
Development Institute (ODI) in London, UK. He
was previously with the CIMMYT Economics
Program, where he worked on training and
methods for on-farm research.
Email: r.tripp@odi.org.uk

Toon van Eijk. Toon is a farming systems agrono-
mist who has worked in Mozambique, Kenya,
Tanzania and Zambia over the last 20 years.
Having recently completed his thesis at
Wageningen University on FSR and Spirituality he
is now a freelance advisor on rural development
based in his home in Dar-es-Salaam, Tanzania.
Email: tvaneijk@ud.co.tz

Abbreviations and Acronyms

AAFSRET African Association of Farming Systems Research Extension and Training
AEZ Agroeconomic zone
AFSA Asian Farming Systems Association
AFSRE Association for Farming Systems Research and Extension
ARFSN Asian Rice Farming Systems Network
ARPT Adaptive Research Planning Team
AVRDC Asian Vegetable Research and Development Center
CARDI Caribbean Agricultural Research and Development Institute
CATIE Tropical Agricultural Center for Research and Training
CGIAR Consultative Group on International Agricultural Research
CIAT International Center for Tropical Agriculture
CID Consortium for International Development
CIDA Canadian International Development Assistance
CIMMYT International Maize and Wheat Improvement Center
CIP International Potato Center
CIRAD Centre International de Recherches Agronomiques pour le Developpement
CONDESAN Consortium for the Sustainable Development of the Andean Ecoregion
DAREP Dryland Applied Research and Extension Project
ECOGEN Ecology, Community Organization and Gender programme (Clark University)
EMBRAPA Institute for Agricultural and Livestock Research (Brazil)
FAO Food and Agriculture Organization of the United Nations
FM/FI Farmer managed/farmer implemented
FPR Farmer participatory research
FPR-E Farmer participatory research and extension
FSR Farming systems research
FSR-E Farming systems research and extension
FSSP Farming systems support project
GIS Geographical information systems
IADS International Agricultural Development Service
IARC International Agricultural Research Centre
ICA Colombian Agricultural Institute
ICARDA International Center for Agricultural Research in the Dry Areas
ICLARM International Center for Living Aquatic Resource Management
ICRA International Centre for Development Oriented Research in Agriculture
ICRAF International Centre for Research on Agroforestry
ICRISAT International Crops Research Institute for the Semi-Arid Tropics

Abbreviations and Acronyms






Institute of Agricultural Science and Technology (Guatemala)
International Development Research Center
International Fund for Agricultural Development
International Farming Systems Association (formerly AFSRE as above)
Intra-Household and FSR-E Case Studies Project
Interamerican Institute for Cooperation on Agriculture
International Institute of Tropical Agriculture
Indigenous knowledge
International Livestock Center for Africa
International Laboratory for Research on Animal Diseases
National Institute for Agricultural Research (France)
Integrated pest management
Institute for Tropical Agronomic Research
Integrated rural development
International Rice Research Institute
International Service for National Agricultural Research
National Agricultural Research and Extension
National Agricultural Research Institute
National Agricultural Research Service
Natural Resources Institute
Natural Resource Management
Overseas Development Administration (now DFID)
On-farm experimentation
On-farm research
Institute Francais de Recherche Scientifique pour le Developpement en
On-station research
Participatory rural appraisal
Production Systems and Natural Resource Management
Production Systems Programme
Recommendation domain
International Farming Systems Research Methodology Network
Latin American Animal Production Systems Research Network
Resource management domain
Researcher managed/farmer implemented
Researcher managed/researcher implemented
Rapid rural appraisal
Remote sensing
Southern African Association of Farming Systems Research and Extension
Southern African Centre for Co-operation in Agricultural and Natural
Resources Research
Department for Research on Agrarian Systems and Development
(within INRA, France)
Sustainable Agriculture and Natural Resource and Environment
Management Collaborative Research Project
Sustainable Agriculture Research and Extension
Swedish International Development Assistance
United States Agency for International Development
West African Farming Systems Research Network
West Africa Rice Development Association
Women in Agricultural Development program (University of Florida)
Women in Rice Farming Systems


Janice Jiggens, Past President of the International Association for
Farming Systems Research and Extension

As President of the Association for Farming
Systems Research and Extension (AFSRE now
the International Farming Systems Association
(IFSA)) when the book was commissioned, I am
delighted to contribute a foreword to this his-
tory of farming systems research (FSR) and its
applications, seeing it as an opportunity to offer
a personal account of my own love affair with
FSR. It mirrors, in many respects, the sequence
of the text itself a retrospective on my own
baptism; a focus on what are, for me, key
aspects of FSR; the progress made and chal-
lenges remaining; and my own perceptions of
some key lessons learned.


I strayed into FSR at the end of the 1970s
when I was working as a social scientist in the
northern and central provinces of Zambia. The
challenge was to find ways to develop tech-
nologies for, and supply services to, impover-
ished small-scale farmers in areas of high male
outmigration. I became fascinated by the
experiments of some women famers to improve
crops of a traditional green leaf vegetable,
grown between the main cereal crop both for
home consumption and sale in the local mar-
ket. The leaves are rich in minerals, dry well
and form an important seasonal additive to the
relish which accompanies the starchy main
meal, as well as providing cash for household

necessities. But when I persuaded an agrono-
mist from the provincial research station to
visit the farmers to learn more about their
experiments and perhaps give some advice, we
ended up in a blazing argument about wasting
his time just to show me some women growing
weeds! I was forced to think deeply about the
specific value of the vegetables to crop produc-
tion and food systems, about the gender-spe-
cific roles of men and women, and about the
nature of a science-based training in agricul-
ture which could so easily set aside farmers'
knowledge and a crop that was essential to the
livelihood of the women and to the nutrition of
their families.
A second formative experience brought into
question agricultural survey research methods.
My team had developed a questionnaire in the
local language to prepare a statistical sample of
households in an area of shifting cultivation.
However, after a few days in the field I realized
that my male Zambian colleagues were estab-
lishing less formal relations with the women in
the village than that of interviewer and respon-
dent. It clearly did not make any communica-
tive sense to turn up a few hours later with a
questionnaire in hand. Yet the long, drawn-out
methods of the anthropologist were not practi-
cal: what tools and techniques could we use in
the 3 weeks we had to ensure some reliable
degree of rigour and representivity yet were
based on a more natural process of enquiry?
With hindsight, I wish I had paid more attention


to the refinement of the concept of 'recommen-
dation domains' and methods of informal
survey that Mike Collinson and his colleagues
were applying at the time in the central
Much of my field work in Zambia turned
into an exploration of alternative methods,
culled from whatever source book or experi-
enced person then available to me. But it did not
feel like 'good' research. I was learning more
than I had ever done before, but how could I
present this knowledge in a way that would
convince my own peer group?
Towards the end of my stay in Zambia
Robert Chambers and I worked together on a
Basic Needs mission sponsored by the
International Labour Organization. The long
trek up to the shores of Lake Bangweulu gave
us ample time for discussion of these questions,
which Robert himself was also pondering and
exploring, along with many others, as I later
realized. A hazardous canoe trip across the lake
brought us to the old 'goat woman'. She
remains in my memory as our tutor in what
later became known as participatory appraisal
methodology. We worked with her for a day
using techniques still regarded as innovative,
analysing the management of her goats which
were renowned for their twins and good health,
and which she sustained through the careful
recycling of waste and the use of traditional
herbs she grew herself.
These formative experiences added in a small
way to the river of accomplishment docu-
mented in this book by bringing together FSR
perspectives, gender analysis and participatory


As the experiences of researchers around the
world during the 1980s demonstrated, there is
much to be gained by marrying these three
ways of learning and cooperating. On gender
analysis Feldstein and Jiggins1 concluded that
using gender as a focus resulted in a better
description of the system as a whole and
opened the door to a greater understanding of

the opportunities to technical innovation.
Gender adds a little complexity for a lot of
insight, while participatory process and tech-
niques enable farming systems researchers to
engage more effectively with members of farm-
ing communities.
The marriage of FSR-E, gender analysis and
participatory methods has, to a considerable
extent, become common practice. Four
strengths stand out. First, the quality of the
information is better because it is richer, more
deeply contextualized and yet amenable to
aggregation. It is focused yet cost-effective
across scale, where 'scale' is understood as a
recommendation domain. Second, in combina-
tion they can lead to the rapid discovery of con-
tradictions such as the points where experience
diverges, where information is inconsistent and
where interpretations vary. Where there is con-
vergence, consistency and agreement, one can
proceed with confidence along well-established
pathways; where there are contradictions,
assumptions are challenged and further investi-
gation is required. This is the opportunity for
genuinely new theoretical and practical knowl-
edge to emerge. Review of experience suggests
that the combination of FSR-E plus gender
analysis plus participatory methods, prompts
discovery by offering three different 'windows'
into complex situations2. Third, the combina-
tion of perspectives and methods focuses atten-
tion on constraints and opportunities, rather
than problems. In my view, the emphasis on
problems in agricultural research has been a
hindrance to development, if only because it
provides such poor inspiration for effort and for
specification of the potential for change in agri-
cultural reality. Finally, the application of these
methods has drawn attention to the important
and necessary technology-led gains that can be
achieved with poor people living in variable,
diverse and uncertain environments.
However, the combination does have a num-
ber of weaknesses. At the theoretical level,
thinking about systems does not have to be sys-
temic to be useful. But at the practical level, if
the research and technology development objec-
tive is in some way to change the system, then
the methodological toolbox must include the
tools of researching farming as an human activ-
ity. Best practice research is generating a rich and
constructive case book of the participatory

xiv Foreword

methodologies essential to systemic change.
More commonly however, these methodologies
seem to be applied mechanisticially or in an
extractive manner, giving rise to failures in the
change process3. While lip-service might be
paid in research proposals to the role of women
in farming systems, the sad fact is that this
remains a male-dominated area and FSR is still
failing in the proper handling of this essential
There is a third area in which FSR-E practice
falls short of its potential, perhaps because of its
strong historical roots in farm management
economics. Research has highlighted the extent
to which an accomplished end-of-season system
'design' is the desired outcome of responses to
events unfolding through the season. Wherever
the degree of uncertainty is high, the tendency
to assess farming in terms of performance is
particularly marked4, but this tendency is also
to be found in more highly controlled produc-
tion environments5. Given the importance to
resource-poor farmers of managing uncer-
tainty, greater attention should be paid to the
overall implications of dryland farming.
Best practice points the way, for example
through examination of strategies for coping
with varying seasonal conditions and the rules
which guide farming choices. Cox et al.6 con-
ducted elegant research among dryland wheat
farmers in northern Queensland which reveals
much about the nature of contingent decision
making in conditions of uncertainty. They
found decisions to be based on a rather small
number of simple rule sets which were: nested;
triggered by events; interconnected; linked to
additional sets, stable in response to stress (such
as prolonged drought); adaptive to long-term
trends in system states; interpretative; and sup-
portive of simultaneous management of multi-
ple indicators of system performance.
A focus on the management of uncertainty
also suggests a need for greater emphasis in
FSR-E practice on collaboration between farm-
ers and scientists7. Best practice has, in fact,
already moved in this direction, a movement
reinforced by emerging concerns about the rela-
tion between on-farm developments and land-
scape scale resource management. FSR-E is now
being challenged to investigate the relationships
among on-farm systems development, ecologi-
cal systems management and agricultural pol-

icy effects8; and apply participatory applied
research at farm and community levels to nat-
ural resource management.
A final problem lies in the field of FSR-E edu-
cation. For many years I shared the frustrations
of field personnel in trying to turn the human
products of specialist university degrees into
systems thinkers with at least some competence
in working with farmers on system develop-
ment. My early efforts at the University of
Guelph in Ontario to take the lessons of the field
back into academia to produce a generation of
professionals competent in FSR were positive at
the human level. The students reacted enthusi-
astically to participatory methods, interdiscipli-
nary learning and systems thinking. But,
despite the goodwill and support of key individ-
uals, undoubted barriers remained in the rigidi-
ties of departmental structures, the defence of
intellectual territory and the problems of recon-
ciling systems-oriented courses and the needs of
students within the existing study programme.
At Guelph, many of the difficulties of rigidity
between departments have been eased by the
recent creation of an interdisciplinary PhD
offered through a new Faculty of
Environmental Design and Rural Development.
As one who is directly involved as a new-
comer to university life at the Swedish
University of Agricultural Sciences. what most
strikes me is the irrelevance of much of what is
on offer at universities. Many students respond
by finding their own pathways of learning
through ad hoc self-study reading groups and by
making off-campus links to community- and
farmer-based action. The regular programme is
what they have to do to qualify, not what they
want to do to learn. Meanwhile, collaborative
initiatives among coalitions of those with a per-
sonal commitment to change processes are cre-
ating new institutional structures and networks
which bypass existing structures9.
It is encouraging to find that even in the
financially hard-pressed educational environ-
ment of eastern and southern Africa, such
innovations are occurring. For example, a con-
sortia of non-government organizations whose
activities focus on various forms of ecological
farming in partnership with farmers and in col-
laboration with the University of Zimbabwe,
have now developed a degree course which sup-
plements classroom study informed by systems


thinking with periods of field work with the par-
ticipating NGOs.


Learning Together, by Hagmann, Murwira and
Chuma in 199610, documents the development
and extension of soil and water conservation
technologies in Masvingo and Chivi, Zimbabwe.
This example of a new approach was called
kuturaya (to try) by the farmers a translation
of 'research' into Shona. It was based on dia-
logue, on farmers' own real time, on whole-sys-
tem experiments and on a strengthening of
self-organizational capacity at community level.
After two seasons each participating farmer,
besides tied ridging, had at least two other trials
ongoing, selected from among experiments sug-
gested by project staff, local research stations
and farmer innovators, or arising out of discus-
sion of farmers' indigenous knowledge. More
than 10 options have emerged from this joint
process, including mechanical, agronomic, bio-
logical and water saving/irrigation methods
and technologies. Within three seasons from
1992-93, at least 80% of the total of 1136
households within one administrative unit in
Chivi District were practising soil and water
conservation. The important lessons include
the need to focus on integrated land husbandry
since individual techniques cannot overcome
the diversity of conditions nor alone generate
sufficient economic benefit, the value of farmer
involvement right from the start in extending,
enriching and validating the portfolio of experi-
mentation and emerging options, and the
necessity of supporting institutional and orga-
nizational development (within communities
but also within research and extension agen-
cies) in order to support participatory process.
Learning to Learn with Farmers, by Hamilton
in 199511, focuses on a project in southern
Queensland. This provided invaluable input into
research on the development, use and effects of
providing farmers with better tools for monitor-
ing and interpreting system states and trends, as
the basis for informed decision making with
regard to fallow management. The project was
based in a region where 1.8 million ha of the

total cultivated area of 2 million ha was desig-
nated as a 'needs protection' area in the face of
widespread soil erosion. In the space of 4 years,
the interventions raised the percentage of dry-
land wheat farmers in the vulnerable areas who
had adopted one or more fallow management
practice from 30% to 75% some 1600 farmers.
This success was the more remarkable for being
achieved through a period of deepening drought
and economic hardship. An interdisciplinary
team of scientists and extension advisers worked
with farmers on joint systems analysis, and
through periods of sometimes painful and con-
flictual reflection on what was being learned and
how the learning process was occurring. A series
of tools were devised, again largely in collabora-
tion with farmers, to enhance individual and
shared learning about systems states and perfor-
mance. These included: a rainfall simulator, a
soil corer, How Wet (a computer-aided decision
support tool), the Fallow Management Game
(which allows players to expand on and inter-
rogate scenarios generated by the use of the
other three tools) and With and Without (a user
friendly comparative economic analysis tool).
Three lessons stand out: the importance of pay-
ing explicit attention to FSR processes, the power
of stimulating shared knowledge creation and
the need for science leaders and policy makers to
accept that the process will not lead to adoption
of uniform or standardized resolutions across an
ecosystem. Rather, a mosaic emerges adapted to
the systemic requirements at unit levels (the
farm, field and crop).


Despite the growing number of examples of
good practice with demonstrably cost-effective
results, there is much still to be learned at the
cutting edge of FSR-E and a continuing need for
vigilant quality control in everyday practice.
However, to end on a pessimistic note would
give a false picture of the contribution that sys-
tems research in agriculture and resource man-
agement is making to the resolution of urgent
human problems. In my experience, it is a field
of endeavour that attracts dedicated scientists,
researchers and development workers of excep-
tionally high calibre broadly united in a com-
mitment to the betterment of human existence


and the life systems which support it. In the
inclusive direction in which it is evolving, FSR-E
provides a framework for understanding, and
the processes and tools for pursuing the agenda
for human survival captured by Goethe, who
might be regarded as an early member of the
FSR-E family, in the following stanza:
Es ist nicht genug zu wissen
Man muss es auch anwenden;
Es ist nicht genug zu wollen
Man muss es auch tun.


In 1992 I was honoured to be elected as
President of the IAFSRE. One of my main tasks
during my term as President, apart from a per-
manent struggle with financing, was the orga-
nization of the 14th International Symposium
in Montpellier, France, alongside our French
hosts. One issue had been taxing the Board of
the Association and its members since 1989 -
the writing of a history of the Association, and
perhaps a history of FSR in general. FAO, in the

person of Karl Friedrich, then Head of the Farm
Management and Production Economics
Branch, had offered support for the history
within the context of FAO promotion of an FSR-
based approach to development, but possible
authors and editors were all were too busy
'FSR-ing' to take on the job. Then, in December
1994, at Montpellier, it all came together. Karl
Friedrich and I met with Mike Collinson.
Although an FSR veteran and enthusiast,
Mike's commitments over the last 10 years had
inhibited his involvement in AFSRE and he was
attending only his third or fourth (he can't
remember!) symposium of the 14 that had been
held. Now, however, he was due to retire and he
committed himself to the compilation and edit-
ing of an history of FSR on his retirement. He
finally retired in early 1996 and has devoted
much of his time since to finding contributors
and to coaxing their contributions from them.
This is the result 40 contributions from 50 of
the world's leading professionals, from some 20
countries an inclusive sweep of the spectrum
of professions and continents involved in FSR-E.

1. Feldstein, H. & J. Jiggins (Eds), 1994. Tools for the Field. Gender Issues in Farming Systems Research
and Extension. West Hartford, Kumarian Press.
2. Jiggins, J. & K. Raman, 1994. Decennial Review, 1984-1994. Eastern India Farming Systems Research
Programme. New Delhi, The Ford Foundation.
3. Bawden, R., 1995. On the Systems Dimension in FSR. Journal of Farming Systems Research and
Extension, 5(2), 1-18.
4. Richards, P., 1985. Indigenous Agricultural Revolution. Hutchinson, London.
de Steenhuijsen Piters, B., 1995. Diversity of Fields and Farmers. Explaining Yield Variations in
Northern Cameroon. Published PhD thesis, Wageningen, Agricultural University.
5. Leeuwis, C., 1993. Of Computers, Myths and Modelling. The Social Construction of Diversity.
Knowledge, Information and Communication Technologies in Dutch Horticulture and Agricultural
Extension. Published PhD thesis, Wageningen, Agricultural University.
6. Cox, P.G., A.D. Shulman, P.E. Ridge, M. Foale & A.L. Garside, 1995. An interrogative approach to systems
diagnosis: an invitation to the dance. Journal of Farming Systems Research and Extension, 5(2). 67-83.
7. Sperling, L. & P. Berkowitz, 1994. Partners in Selection. Bean Breeders and Women Bean Experts in
Rwanda. CGIAR, Washington, DC.
Sperling, L. & U. Schiedegger, 1995. Participatory Selection of Beans in Rwanda: Results. Methods and
Institutional Issues. Gatekeeper series no. 51, IIED, London.
8. Caldwell, J. & E. Akobundu, 1997. Agricultural Systems and Policy. Blacksburg, Virginia Polytechnic
Institute and State University for AFSRE.
9. Jiggins, J. & D. Gibbons, 1997. What does interdisciplinary mean? Experiences from SLU. Paper pre-
sented to Session 5: Agricultural Knowledge and Information Systems. 13th European Seminar on
Extension Education, 1-6 September. University of Dublin, Dublin.
10. Hagmann, J., K. Murwira & E. Chuma, 1996. Learning together: development and extension of soil
and water conservation in Zimbabwe. Quarterly Journal of International Agriculture, 35(2), 1-14.
11. Hamilton, N.A., 1995. Learning to Learn with Farmers. Published PhD thesis, Wageningen. Agricultural

Chapter 1


Mike Collinson

Even within the choir of advocates there has
long been controversy on terminology in farm-
ing systems research (FSR)1. It raised its head
again during the preparation of this book. I
hope I have outflanked the controversy by refer-
ring to FSR and its applications. FSR itself is
defined as a diagnostic process; a basket of
methods for researchers to elicit a better under-
standing of farm households, family decisions
and decision-making processes. Its applications
use this understanding to increase the effi-
ciency in the use of human and budgetary
resources for agricultural development, includ-
ing research, extension and policy formulation.
These are important applications, both for those
countries which rely on the traditional agricul-
tural sector to drive their economic develop-
ment, and for other countries where that sector
is small in terms of population, but where a
social conscience demands measures to combat
rural poverty.
I have tried to give the book diversity
through the number and origins of its contribu-
tors, and coherence through its structure.
While the application of FSR in developed coun-
try agriculture is occasionally illustrated, the
book is primarily focused on FSR in its original
role, with small, resource-poor farmers in devel-
oping countries. The origins of contributors are
sometimes deceptive. Europeans and North
CAB International 2000. A History of Farming
Systems Research (ed. M. Collinson)

Americans write about experiences in Africa,
Asia and Latin America, for expatriates indeed
dominate the early history of FSR, itself per-
haps a factor in the resistance to change in
institutions in many developing countries. An
expanding professional capacity there began to
make itself felt in FSR's application and evolu-
tion in the 1980s, yet institutional change is
still perhaps the single biggest constraint to
wider application. Similarly, the early days of
FSR are male dominated but the number of
contributions in the book from women demon-
strates how they have increasingly asserted
themselves in agricultural development.
The book is divided into five parts (each with
an editorial introduction) and 12 chapters, each
with several contributors. Part I of the book tries
to capture the origins and the essence of FSR; its
conceptual framework and some of the methods
central to the understanding of the farming of
resource-poor communities. It begins with con-
tributions from a group of pioneers fondly
labelled 'the old dogs'. Part II examines the appli-
cation of FSR understanding to the choice and
development of technology, to the planning and
evaluation of extension, and to policy formula-
tion. Part III focuses on efforts made to incorpo-
rate FSR into agricultural research and
extension systems in Africa, Asia and Latin
America. It also covers the essential companion
to institutionalization; the training of profes-
sionals in FSR. Part IV looks at the organization
of FSR professionals, with contributions on the

Chapter 1

growth of associations and networks in Africa,
Asia and Latin America, as well as on the
Association for Farming Systems Research and
Extension (AFSRE), subsequently renamed the
International Farming Systems Association
(IFSA). These accounts are complemented by
commentaries from professionals in agronomy,
farm management and rural sociology on the
interaction of these disciplines with FSR. The
fifth and final part of the book turns to the
future. Current practitioners discuss cutting
edge methods and applications in FSR and the
final chapter looks at the lessons of the past and
the possibilities for the future. It sets out how
FSR has moved toward its original goal a better
understanding of small farmers and, as sys-
tems applications in agriculture proliferate, asks
whether it still has a distinct role. The editorial
introductions to each of the five parts outline
the contributions and offer a personal commen-
tary on the theme covered. Where appropriate,
this summarizes the evolution of that theme,
highlighting both progress and unresolved
issues. Three unresolved issues pervade the edi-
torial introductions and take centre stage in
Chapter 12; the scope of FSR, its place in the R &
D process, and strategy for institutional change.

FSR was one of a number of threads from sys-
tems thinking that reached into agricultural R
& D in the late 1960s and early 1970s. Crop
modelling, dominated by the disciplines of
physiology and agronomy, was another innov-
ative thread, as was cropping systems research,
recalled by Dick Harwood in Chapter 2 as
underpinning the origins of FSR in Asia.
Eagerly grasped by a variety of constituencies,
the early, tight focus of FSR rapidly widened.
Texts on systems and agricultural develop-
ment, including those by Penning de Vries,
Teng and Metselaar in 1993, Dent and
Macgregor in 1994 and CIRAD in 19962,
demonstrate the growing range of systems
applications in agriculture. It has become
unclear, perhaps even confusing, to practition-
ers, how FSR is best viewed within that spec-
trum. Proliferating constituencies for systems
applications in agriculture, and confusion over
the scope of FSR have arguably distracted from
its practice and institutionalization.

FSR was an innovation in the research
process, emerging from field practitioners, an
early effort to bridge the gap between the needs
and capacities of small, resource-poor farmers
and publicly funded agricultural research
establishments. Early in the book, founder
members of the FSR family talk about its ori-
gins. The common threads through the differ-
ent accounts leave no doubt that in the 1960s
and early 1970s the same problem was widely
identified across the developing world; tech-
nologies recommended as a result of agricul-
tural research investments were, in general,
inappropriate to the priorities and circum-
stances of small farmers. Field practitioners
recognized the importance of the problem and
targeted a better understanding of small farm-
ers and the way they make decisions, as a path
to its solution. Their concern for appropriate
improvements for small-scale, illiterate and
resource-poor farmers was the origin of FSR
and remains its foundation.
But FSR has also been elaborated, and for
some confounded, by the scrutiny of academics.
Development theorists, often economists, have
criticized the narrowness of conceptual frame-
works pinned together by practitioners preoccu-
pied by technology adoption. These originally
ignored such issues as intra-household equity,
population dynamics, intergenerational equity
and sustainability, and the wider macro and pol-
icy linkages that these imply. 'Imported' meth-
ods, driven mainly by academics doing research
to add to theory, or to test out methods in new
circumstances, have sometimes diverted profes-
sional attention from the operational circum-
stances of developing countries, the modest
institutional capacities and thin budgets with
which FSR professionals were wrestling. A noto-
rious example in farm management was the
quest to apply linear programming to the small-
farm sector in the 1960s. Promoted by the 'have
tool will travel' brigade, usually from academia
in the USA, it has not yet made a significant
operational impact in developing country agri-
culture. Its failure has been due to the intensive
data collection efforts required, and the very
high costs of bringing the results of program-
ming to bear on farm units with such low levels
of income that even major improvement would
offer little return for the costs of the research
and advisory process.


Still today, a generation on, in many of the
countries where the small-farm sector remains
crucial to both the national economy and to the
environment, the research/farmer interface
remains a critically weak link in the develop-
ment process. Thus, despite a 25-year history,
FSR remains an innovative component in the
process for agricultural R & D. The prolonged
gestation for FSR reflects the forces governing
innovation particularly innovation in public
institutions in developing countries, and is
itself a lesson for both governments and aid
agencies. There has been great difficulty in fit-
ting FSR into agricultural institutions. Is this a
failing in FSR as an innovation, or are the
power dynamics and the entrenched institu-
tional and professional interests in national
agricultural R & D too formidable for change?
Has the timing of its introduction been inappro-
priate? The book examines these important
ongoing issues. Indeed, the history of FSR is a
case study of the dynamics of institutional
innovation in developing countries.
The introduction of FSR has been compli-
cated by:

* The need for changes in professional atti-
tudes and institutional orientation and orga-
* The biases of the inherited, often colonial,
establishments, in both agricultural educa-
tion, research and development; expatriate-
driven, Western mind-sets, isolated from the
small-farm sector, with inappropriate
criteria for success.
* Differences between commercial farmers,
often driving public programmes in many
developing countries, and resource-poor

Small farmers do not behave like commercial
farmers. They are not organized to interact with
the wider market economy, nor are they politi-
cally articulate like commercial farmers. These
had attracted a set of service institutions, for
example in credit and insurance, for protection
against the vagaries of weather and the market.
These older institutional processes, oriented to
and organized for large farmers, cannot operate
cost-effectively with small farmers who, in the

absence of an appropriate enabling infrastruc-
ture, must manage their environment directly by
their own decisions and by their activities both
on and off the farm. Small farmers often cannot
use the technologies appropriate for commercial
farmers and always need explicit consideration
in agricultural R & D. These insights have given
rise to the development of new investigative
methods to manage the different circumstances
of resource-poor farmers under conditions of
scarce professional and financial resources. A
start has been made in reorganizing agricultural
R & D institutions to implement the new meth-
ods and to adjust higher agricultural education
to achieve congruity between the mind-sets of
peasant farmers and professionals to encourage
mutual respect and partnership in agricultural
A parallel feature of the last 15 years, and
one which holds great hope for the future, has
been the growth of FSR professional associa-
tions. FSR associations attract people from a
range of disciplines, from agronomy, ecology
and plant breeding to economics, anthropology
and rural sociology. The growth of these pio-
neering associations has received much of its
impetus from the leadership of university profes-
sionals, who established an annual symposium
for FSR-E in the USA in the early 1980s. This
evolved into the AFSRE and associations and
institutional networks now exist at the continen-
tal level in the USA and Asia, and at the regional
level in Africa, Latin America and Europe.
Several contributions to this book document the
evolution of these associations which promote
interdisciplinary interaction around key prob-
lems, encourage independence for professionals
in developing countries and complement alle-
giance to discipline with allegiance to people in a
refocusing of the R & D process in agriculture. In
Africa, Asia and Latin America FSR associations
are moving professionals out from under the
spell of developed country fora, finding their feet
in their own context, and helping to bring both
education and development processes into line
with the needs of local people. It is good to be
able to record progress towards these goals. But
it is important to record that these gains remain
fragile and there is a danger that governments,
courted by the dynamics of growth at any price,
may despair of their smallholder constituencies
as an engine to achieve it.

Chapter 1

Appropriate intervention for farm improve-
ment remains the heart of FSR. Experience
has widened the portfolio of interventions
beyond the early emphasis on technology
development. Accumulating insights into the
nature of the traditional agricultural sectors
of developing countries have shaped the evolu-
tion of an FSR process for their successful
development and deployment. The early
insights included:
* Recognition that vast numbers of small
farms dominate agricultural sectors in many
developing countries under widely diverse

* Recognition that on one small farm, a major
improvement of productivity, even 100%, is
a small absolute benefit, and costs of achiev-
ing it must be low.
* Recognition that appropriately qualified
agricultural professionals are an extremely
scarce resource.

The scope of FSR and the strategy for promotion
and institutionalization, perhaps the funda-
mental issues of FSR, are revisited in the final
chapter. I hope this book will provide a founda-
tion on which a second, or now perhaps a third,
generation of farm systems practitioners can

1. Merrill-Sands, D., 1986. Farming systems research: clarification of terms and concepts. Experimental
Agriculture, 22, 87-104.
Simmonds, N.W., 1985. Farming Systems Research a Review. World Bank Technical Paper no. 43.
World Bank, Washington D.C.
2. Penning de Vries, E, P. Teng, & K. Metselaar (Eds), 1993. Systems Approaches for Agricultural
Development. Kluwer, Dordrecht, and IRRI, Los Bafios, Philippines.
Dent, J.B & M.J. Macgregor (Eds), 1994. Rural and Farming Systems Analysis European Perspectives.
CAB International, UK.
CIRAD, 1996. Systems Oriented Research in Agriculture and Rural Development. Lectures and Debates
from an International Symposium, Montpellier, France. CIRAD-SAR, Montpellier.

Part I

FSR Understanding Farmers

and Their Farming

Mike Collinson

In my general introduction in Chapter 1, I1 skirted the historical controversy on terminology by dis-
tinguishing the process of farming systems research (FSR) from its applications in technology
development, in extension, and in policy formulation. Part I of the book, in two chapters, deals with
the development of the FSR process in its role of understanding farming systems.

Chapter 2 features personal accounts by five
'old dogs' of the experiences which drew them
into the development and promotion of FSR in
the late 1960s and early 1970s. They recapture
the insights which convinced them that FSR
could improve the relevance of conventional
agricultural research to the situation of count-
less small farmers in the developing world. The
contributions vary from personal, even anecdo-
tal, to semi-formal. Each offers lessons and
many of the issues raised from the 1960s and
19 70s remain issues today.
The contributions are from German Escobar,
working in Latin America, Pete Hildebrand in
central America, Dick Harwood in south-east
Asia, David Norman in west Africa, and myself in
east Africa. German, David and myself are farm
economists, with Dick the only thoroughbred
agronomist. I have to mark Pete down as 'hybrid';
an agricultural economist by training, much of
his best known work has been in the analysis of
stability in biophysical parameters important to
farmers. Some of the most telling points made in
Chapter 2 are listed in Box I.1, all arise in more
than one account, and some arise in all five.
Even in the early days linkages were impor-
tant and some of these commonalities can be
attributed to the interactions that occurred
across continents. German Escobar mentions
the influence of Hans Ruthenburg1, another
'old dog', and both he and Dick Harwood
acknowledge the value of interacting with
David Norman in formulating an approach to

CAB International 2000. A History of Farming
Systems Research (ed. M. Collinson)

the problem of non-adoption by small farmers.
Yet much was clearly spontaneous. All five 'old
dogs' overlapped in the timing of their 'conver-
sions' to FSR, all were focused firmly on the
small resource-poor farmer and agricultural
technology. It is as though the 'bones' of the
process were buried around the world but were
dug up by the 'old dogs' from one continent to
the other, sometimes in a different sequence.
Chapter 3 also has five contributions. These
address the conceptual framework and four
aspects of methodology I judged as central to an
in-depth understanding of farm systems: typol-
ogy or characterization, diagnostic methods,
gender analysis as a neglected dimension of
diagnostics, and the crucial step from problem
identification to an understanding of the
causes. Robert Hart analyses the evolution of
the conceptual framework for FSR. My contri-
bution on typologies highlights how new meth-
ods for reconciling physical and human
attributes are contributing to a revolution in
typologies for agricultural development. John
Farrington reviews the evolution of diagnostic
methods and asks how far farmer participatory
research (FPR), as a less extractive, more colle-
gial approach, is a development of FSR or an
alternative to it. Hilary Feldstein describes the
gradual acceptance of gender analysis, a diag-
nostic method to capture the understanding of
gender roles in the household and on the farm.
Finally, Robert Tripp, a colleague with the eco-
nomics staff of the International Maize and
Wheat Improvement Center (CIMMYT) in the
1970s and 1980s, delves into causal analysis, a

Part I

step in the diagnostic process to help ensure
that on-farm experimentation, the most expen-
sive step in the on-farm research (OFR)
sequence, is attacking the most relevant issues
in the most appropriate way.
None of the contributions here is a manual
on 'how to do it', there are plenty of these. Each
is a commentary on the evolution of ideas in
these defined areas of the FSR process, aimed
towards the understanding and improvement of
small-farmers' systems.


Perhaps the most telling change over time is the
shift from the farm system per se as a framework,
to a hierarchy of systems within which the farm-
ing system is one of a number of levels, a change
pioneered by Robert Hart2. In his contribution
here he uses communities and watersheds as
examples of wider systems in a hierarchy.
One important consequence of the wider
framework is the diversity of perspectives
embraced. Decision makers at all levels in the
hierarchy have different views on a given issue,
shaped by their varied roles: communities may
take a different view from individual farmers on
those on-farm activities with impacts beyond
the farm boundaries externalities in the econ-
omist's jargon. Policy makers need to reconcile
the perspectives of many such communities
while pursuing aggregated interests at the local
or national government levels. The diversity of
perspectives is synonymous with the concept of

multiple stakeholders. It illustrates why FSR has
moved from formal to informal farmer surveys,
to embrace farmer participatory methods,
stakeholder analysis and conflict resolution.
The hierarchy of systems also responds to ear-
lier criticisms of FSR, including its failure to
embrace long-term sustainability and to
address the important interaction with policy.
FSR needs to reconcile the wider perspectives at
higher levels of the hierarchy with the ways in
which it helps farmers improve their welfare.
Again this highlights its scope, an issue taken
up in the concluding chapter of this book.
For those operationalizing such frameworks
over the years the phrase 'the farming systems
approach' has been a valuable 'shorthand'. On
reflection, in some circumstances, the phrase
has perhaps been counterproductive. Certainly
it has been an anathema to the ears of the
research establishment, carrying as it does the
implication that FSR is an alternative to the tra-
ditional process. In serious circles FSR has
always been perceived as a supplement, or com-
plement, to the existing process. It is clearly no
substitute for reductionist experimentation. Yet.
partly because its promotion was sometimes
perceived as confrontational, its role in focusing
and interpreting more formal experimentation,
both on and off the research station, has
remained controversial. Insensitive promotion
may have been one of the most serious hurdles
to institutionalization. Use of the word
'approach' perhaps originated in the early com-
petition between agencies in promoting their

Box 1.1. Common insights from the early days of FSR.
* The limited adoption of research recommendations by small farmers, yet their obvious skills in man-
aging their environment with limited resources
* Priorities in research set from a researcher's perspective, based on the need to get the most out of the
commodity under local biophysical conditions
* The contrasts between farming methods on research stations and on the surrounding small farms, and
between the exalualion criteria used by researchers and by small farmers
* Small farmers' willingness to learn and to change, and the way they make these changes; by trying
out new ideas on a small scale and adapting these, often quite radically, to their own circumstances
* Low professional/farmer ratios and the need to deal with numbers of farmers together
* The farming system as a basic unit for agricultural R & D
* The trade-off between coverage and intensity of effort, and the search for cheaper methods; qualita-
tive versus quantitative investigation to provide understanding
* The difficulties of interdisciplinary research and capacity building in new methods
* The problem of social scientists, often new, junior, recruits, criticizing the products of senior mem-
bers of the established research disciplines with which they need to collaborate
* The value of learning through experience

FSR Understanding Farmers and Their Farming

own paradigms. Although FSR's application to
technology development has four agreed stages
- diagnosis, design, implementation and evalu-
ation there have been contrasting approaches
to its operationalization and practitioners have
attempted to categorize these, as in Norman
and Collinson, 19853. One illustrative contrast
is between the Tropical Agricultural Center for
Research and Training (CATIE) and CIMMYT.
The CATIE 'approach' was labelled 'formal,
quantitative and rigorous' in the early 1980s. It
was characterized by intensive data collection,
parameter measurement and computer model-
ling. Analysis held all options open across the
system and all model parameters were poten-
tially variable embracing the dynamics of sys-
tem evolution and the opportunities for changes
in policy, and avoiding some of the criticisms
made of less sophisticated approaches. In con-
trast the CIMMYT 'approach' was labelled
'informal and qualitative'. This purposefully put
to one side large areas of the system as it
focused in to detail on maize or wheat. It moved
from an understanding of the climate, markets
and policies facing local farmers, through an
understanding of farmers' strategies in manag-
ing their system, to analyse how the specific
practices used to manage the maize or wheat
enterprises were dictated by this wider context.
The CIMMYT approach was pre-focused, clos-
ing down options as understanding grew of
how maize or wheat management was shaped
by its system context. The CATIE/CIMMYT com-
parison also illustrates extremes in the search
for acceptable compromise between theoretical
desirability and operational possibility.
Characterization is an important first step in
the FSR process of description and analysis. In the
current literature the term 'characterization'
often seems to replace 'diagnosis', but properly
alludes only to the description, or profiling, of dis-
crete units, such as the agroecology and the farm
enterprise pattern. It does not, on the whole,
assume an understanding of the farming system.
Many development programmes are still imple-
mented by administrative units, even though it
has been clear since the early 1960s that survey-
ing of any sort using administrative boundaries,
cannot, except serendipitously, differentiate types
of farming. Since surveying across farming
systems confuses description and thoroughly con-
founds understanding, the grouping of farmers

into types by profiles of the systems they operate
takes on particular importance. Essentially a
means of stratification, it seeks to maximize
differences between types and minimize sources
of variation within them. No stratification device
is perfect and the farming system is no exception.
Beyond this, important parameters such as yields
will often express variation more strongly at the
farm rather than the farming system level, parti-
cularly across systems within the same agro-
ecology. From year to year these yield variations
have a major impact on the relative performance
of farmers operating the same system.
Historically, biophysically derived zones have
dominated typing in agricultural R & D, mainly
in the identification of uniform zones of land
and crop potential. While it is true that a rela-
tively narrow set of activities offers the best
physical potential for any area, a much wider
set of economic production opportunities typi-
cally exists for the area. From a purely land use
perspective, the physical performance of this
wider set may be relatively poor, some may even
threaten the integrity of the resource base.
Nevertheless small farmers choose activities
that are economically superior from the wider
set. CIMMYT's 'recommendation domains'
(RDs) of the late 19 70s were a pioneering step
in defining groups of farmers for whom the
same changes would be relevant. Since then
widening acceptance of the link between
poverty and environmental degradation is forc-
ing further reconciliation between traditional
physically based definitions of zones, in terms of
climate and soil, and people-based definitions.
In diagnosis proper, the unravelling of the
complexity of the household has been an impor-
tant step towards better understanding. What
had historically been seen as 'the farmer' was
overtaken, as understanding improved, by the
interaction, and indeed negotiation, between
household members for access to resources and
control over output flows. The credit for unravel-
ling this tapestry and for earning gender
analysis a place in the FSR, goes to a relatively
small group of intrepid women who kept their
eyes, and their actions, firmly on the unacknow-
ledged role of their gender in traditional agricul-
ture. Prominent among the group, both for their
articulation of the issue, and persistence in its
pursuit in the field, are Hilary Sims Feldstein,
Janice Jiggens, Joyce Moock and Susan Poats.


Despite the increasing pervasiveness of mar-
ket forces, non-market objectives putting food
on the table day in day out, with family food
preferences more or less satisfied depending on
the vagaries of the season continue to drive
many of the actions of the majority of small
farmers. The direct management of risk
through the enterprise pattern and farm man-
agement practices, rather than through market
institutions offering overdrafts, credit and
insurance, as they do to commercial farmers,
dominates management strategy. Thus, even in
farming systems where market access offers
cash earning opportunities, subsistence and
survival goals often take priority in the alloca-
tion of family land and labour. I believe there
are still relatively few situations in the tradi-
tional farm sector where market opportunities
are so valuable that 'basic needs' goals and
strategies are subordinated to their exploitation.
Understanding the specifics of these goals and
strategies for each important farming system is
the key to identifying interventions which are
valuable to farmers, to designing relevant
extension and credit programmes, and to the
sensitive formulation of policy.
The search for cost-effective methods of
gaining understanding led to the development
of rapid rural appraisal, informal surveys in
which representative farm families and farmer
groups are engaged in conversations with
researchers that are allowed to flow, often
guided by the farmers. Initial conversations in
the process are descriptive, later ones, usually
with other families, are analytical, seeking veri-
fication of hypotheses set up from the earlier
descriptions in order to unlock family priorities
and farmer management strategies4. The evolu-
tion of diagnostic methods for FSR was accom-
panied by strong professional debate at each
stage. Two social science schools emerged, the
hard data modelling school, exemplified by the
CATIE approach within FSR, and more broadly
by researchers using formal economic models
or seeking to contribute to development theory.
The place of modelling within FSR remains an
issue and is discussed further below. The scien-
tific credibility of the informal approach which
emerged in FSR was questioned by natural sci-
entists in agricultural research institutions.
Battle was engaged between natural scientists
critical of the qualitative research methods, and

FSR practitioners horrified by the preoccupa-
tion with precision and apparent unconcern for
relevance in agricultural research. Qualitative
understanding has been legitimized by systems
writers5. Further, occasional rigorous compar-
isons of results from informal and formal sur-
vey work showed no significant distortions of
reality from an informal approach6.
Bridging the gap between diagnosis and
action has long been acknowledged as a
weakness both in FSR and in FPR. Clive
Lightfoot (Chapter 11.5) has termed it the 'so
what' syndrome we have an insightful
description of local farming, but where do we go
from here? I regard causal analysis as another
of the key contributions from the International
Agricultural Research Centres (IARCs). This
came from a collaboration between the CIM-
MYT economics team of the 1980s and CIAT
agronomists7. It is a technique to carry the
process from the diagnosis of problems to the
identification of solutions, in a sense, the ful-
crum of OFR. It may involve experimentation
by natural scientists to determine physical
causes, or research by social scientists to iden-
tify economic and cultural ones. In his contri-
bution Robert Tripp also distinguishes
proximate and ultimate causes, and, while
warning against pursuing the causal chain ad
infinitum, brings out the value of understanding
a number of links to multiply the strategies and
options for solution. It is a technique which I
believe has taken its place among the 'best prac-
tices' of contemporary diagnosis in FSR.

The better understanding of farming systems
has both improved old paths for farm improve-
ment and identified new ones:

* Shaping new technologies to the cultural
circumstances and resource constraints of
the existing farming system.
* Creating a better understanding of develop-
ment dynamics, particularly the watershed
between increasing the area cultivated and
yield intensification as land scarcity grows.
and the benefits of wider market access
allow greater relevance in interventions.
* Reinforcing farmers' own strategies for man-
aging uncertainty and raising productivity.

FSR Understanding Farmers and Their Farming 9

* Seeking greater flexibility of action for farm-
ers in the face of climatic and market uncer-
* Looking beyond the system itself to those
factors influencing farmers decisions about
their system. Breaking resource constraints
by action off the farm, at a higher level in
the hierarchy.
* Revising community or policy decisions to
change farmers' incentives and encourage
practices that reduce resource degradation.
New institutional orientations can be particu-
larly effective where degrading practices are a
response to market and price uncertainty.
* Developing strategies for the introduction of
new market production opportunities.
Three issues persist with the process for under-
standing small-farm systems. The first, much
the most threatening, is superficial diagnosis.
Continuing weakness in current diagnostic
practice jeopardizes the credibility of the
process and inhibits its mobilization in develop-
ing country institutions. Superficial diagnosis is
also closely related to the second issue; whether
sound systems understanding adds value in the
diagnostic process over and above collegial
farmer participation. Both this and the value of
quantitative models in FSR are questions cloud-
ing a conclusion on best practice.

Superficial diagnosis
In the early days of FSR some agronomists
moving off-station into on-farm experimenta-
tion simply brought their old programme of sta-
tion-based experiments with them, ignoring
diagnosis. Diagnosis was also often decried by
extension and development professionals who
insisted that they knew all there was to know
about their farmers. It is a tendency which has
persisted and one which is encouraged by
superficial problem identification from FSR
teams; 'low yields', 'poor soil fertility' or 'crop
disease' are scarcely insightful. Poor training,
partly a result of the rush to climb aboard the
FSR bandwagon or perhaps gravy train of
the early 1980s, has been responsible and the
fact that there are still relatively few profession-
als aware of what qualitative diagnosis can pro-
vide is a reflection of the slow pace of change in
university curricula. Good diagnosis depends on
a sound grasp of the principles of agronomy,

production economics and farm management,
and on disciplinary interaction.
It is worth looking briefly at the sort of
insights which emerge from good diagnosis.
Examination of background information on cli-
mate and soils, markets and institutions is an
essential basis. It builds up an understanding of
what local farmers must manage, and helps
identify the facets of the farmers' environment
that create problems. Risk management plays a
large role in resource-poor farmers' decisions
and is a key area for diagnosis. A good example
is within season rainfall variability in dryland
agriculture. In southern Zimbabwe, CIMMYT
collated rainfall data for a 30-year period by
pentads8. Analysis identified three patterns of
within season drought: some 30% of years
showed a significant delay in the onset of rains,
some 30% showed an early finish, and some
40% a 3-week mid-season drought occurring
anytime within the 3-month period from early
December to the end of February. Some years
showed more than one type of drought, others
showed none.
The main starch staple for the system was
maize. The analysis of rainfall data provided a
sound basis for discussion with farmers' on how
they managed the uncertainties these patterns
created. Losses from drought were a well under-
stood risk in local farming and farmers
responses on the occurrence of the three types
of drought were related back to the rainfall data
for verification. Farmers reported making a
series of two, three, four or more plantings over
the period November to February, increasing
the chances that one or more plantings would
not be caught by drought in a critical period of
growth. Later plantings were made with early
maturing varieties. Those early plantings criti-
cally damaged by delayed onset, or by early mid-
season drought, were replanted with earlier
maturing material. These demonstrate both
pre-emptive and reactive risk management
strategies. Farmers routinely make several
plantings with maizes of differing maturity
period in anticipation of drought. Replanting,
on the other hand, is a reaction to drought
occurrence. These strategies benefited from the
policy of government purchase of maize at a
fixed price, common to many African countries
until the late 1980s. Farmers could not over-
produce maize, it was a win-win situation for

Part I

them. Strengthening their strategies for
drought management in maize, both pre-emp-
tive and reactive, offered an important focus for
an improvement programme, an importance
enhanced by the opening of markets under
structural adjustment initiatives.
A vital thread running through the diagnos-
tic process is the interaction between biological
and social scientists. Interdisciplinarity is well
established between breeders, pathologists,
physiologists and agronomists in classic agri-
cultural research the new element in FSR is
the social scientist. One early rendering of the
interaction ran like this:
The biologist brings to the diagnostic process a per-
ception of the ideal technical management for
crops in the conditions of climate and soil under
which farmers are operating. The social scientist
brings an understanding of farmers' priorities and
the constraints operating on them, limiting the
ways in which they can adjust their management.
The biologist evaluates the background informa-
tion on natural conditions to assess crop potential
and management practices likely to be important.
The social scientist evaluates background informa-
tion on economic, cultural and institutional condi-
tions. The biologist learns about farmer
management practices and identifies changes
which would better exploit biological potential. The
economist learns why farmers are doing things the
way they are and identifies when resources are
available and when limiting. In interpreting the
survey work the biologist puts forward ideal
changes from a technical point of view and esti-
mates the likely improvement in yield. The social
scientist assesses their possible profitability and
their compatibility with the ways farmers currently
allocate their resources to realise family priorities9.

Although they remain 'offstage' here, interac-
tion with farmers is central to the process.

Participation versus systems understanding
Public attention and much funding has shifted
from FSR to participatory research in which
farmers are seen as full, indeed dominant part-
ners. Most of us follow Robert Chambers in
advocating empowerment for small-scale farm-
ers10. Empowerment, however, has an affinity
with community development. Both concepts
are abstract and require operational goals and
processes beyond themselves for their imple-
mentation. Technology generation includes just

such a process and was adopted as one vehicle
for implementing the participatory concept.
FSR best practice has absorbed many methods
from the participation portfolio. However, at the
extreme it can be argued that much of the
process adopted as a vehicle for empowerment
by participators has been usurped by them.
There is an insistence that farmers and commu-
nities make the decisions and that agencies ser-
vice these. Like early FSR this seems to neglect
the need to reconcile local, national and indeed
global interests. It also seems, rather like the
extreme advocates of indigenous knowledge, to
deny that outside knowledge will have a key
part to play in providing sustainable livelihoods.
Recently cracks have appeared between rhetoric
and reality in participation. Some were high-
lighted by Rhoades in 1998: 'the social scientist
who attempts to raise analytical points about
stratification, differential access to power and
resources, and other social shaping dynamics
are accused of being top-down and then are
marginalised by turf guarding NGOs'"1. It is an
issue to which we return in further editorials
and in the final chapter.

Quantitative modelling as a diagnostic tool

Although the value of qualitative understand-
ing is now widely accepted, modellers continue
to press their case for the use of quantitative
socioeconomic models within an FSR context.
Against my own history of moving away from
quantitative to informal surveys to reduce costs
and gain coverage in terms of the systems
researched for a given tranche of manpower
and budget, this almost seems sacrilegious.
Modelling apart, experience shows there are a
number of valuable roles that hard data can
play: verification of informal survey conclu-
sions, including quantification of the incidence
of farmer priorities and strategies across the
population. Evidence of impact is valuable, even
the simple recording of adopters over time.
However, as Roberto Quiroz and his colleagues
demonstrate in Chapter 11.3, and as Dent and
others have long argued. models can make a
valuable contribution to understanding, partic-
ularly in ex ante evaluation of the impact of
interventions on the existing system. Questions
inevitably arise: can modelling be reconciled
with a low cost approach to small-farm

FSR Understanding Farmers and Their Farming 11

improvement still necessary in many develop-
ing countries where professional manpower
remains limited? With what specifications of
objective function does formal modelling
improve on a sound qualitative understanding
of the system? Is the degree of improvement
worth the extra cost and effort of intensive data
My own conclusion is that the application of
quantitative modelling has advanced a great
deal since the 1970s, particularly in the ease of
analysis. The data requirements, however,
remain heavy and expensive for local application
in an FSR context where professional manpower
is scarce and budgets low. Also, for those small-
farm systems still dominated by subsistence pro-
duction, representation of the objective function
remains weak. For me it is too 'hands off' and

has too many pitfalls to advocate its incorpora-
tion in routine FSR diagnosis. That said, it is
important to recall that many early practitioners
had a farming systems perspective imbued
through an involvement in socioeconomic mod-
elling. The important problem of superficial
diagnosis in current FSR practice suggests mod-
elling might have a valuable role in training. As
establishments mature, the idea of using the
staff deployed in the field for experimentation to
collect the detailed data for modelling is attrac-
tive. The goal might be a portfolio of models for
major farming systems for training in both farm
improvement and policy analysis, and for policy
analysis itself. There is a good case for universi-
ties developing and using locally relevant models
to ingrain a farming systems perspective into
their agricultural students.

1. Ruthenburg, H., 1980. Farming Systems in the Tropics, 3rd edn. OUP Oxford.
2. Hart, R.D., 1982. An ecological conceptual framework for agricultural research and development. In:
Shaner, WW, Phillipp, EP. & Schmehl, WR. (Eds) Readings in Farming Systems Research and
Development. Westview Press, Boulder, Colorado, pp. 44-58.
3. Norman, D. & M. Collinson, 1985. Farming systems approach to research in theory and practice. In:
Remenyi, J.V. (Ed.) Agricultural Systems Research for Developing Countries. ACIAR, Canberra.
4. Collinson, M., 1974. Transferring technology to developing economies: the example of applying farm
management economics in traditional African agriculture. World Development, 2(2), 93-7.
Hildebrand, P., 1976. Generating technology for traditional farmers: a multidisciplinary methodology.
ICTA, Guatemala City.
Rhoades, R., 1981. The Art of the Informal Agricultural Survey. Training Doc. 1982-2, Social Science
Department, CIP, Lima.
5. Checkland, P.B., 1981. Systems Thinking, Systems Practice. John Wiley, Chichester.
6. Franzel, S. & E. Crawford, 1987. Comparing formal and informal survey techniques for farming sys-
tems research: a case-study from Kenya. Agricultural Administration and Extension, 27, 13-33.
7. Tripp, R. & J. Woolley, 1989. The Planning Stage of On-Farm Research. Identifying Factors for
Experimentation. CIMMYT and CIAT, Mexico, D.E and Cali.
8. CIMMYT & Government of Zimbabwe, 1981. Report no. 5. Demonstrations of an Interdisciplinary
Approach to Planning Adaptive Agricultural Research Programmes. CIMMYT, Nairobi.
9. Collinson, M., 1981. A low cost approach to understanding small farmers. Agricultural Administration,
10. Chambers, R., 1994. The origin and practice of Participatory Rural Appraisal. World Development, 22
(7), 953-69.
11. Rhoades, R., 1998. Participatory Watershed Research and Management; Where the Shadow Falls.
Gatekeeper series no. 81. IIED, London.

Chapter 2

FSR: Origins and Perspectives

German Escobar

Neglect of the peasant sector brought an emphasis on the use of FSR methods in the field by organizations
with a social conscience. As a young professional 22 years old, as you can imagine, I found the opportunity
to share my theoretical training and political ideas with very small farmers living in poor conditions very

2.1.1 Introduction
The introduction of the farming system
research (FSR) approach in Latin America and
the Caribbean (LAC) was not an organized,
monolithic event, but a rather haphazard,
serendipitous process. It was introduced as a
part of a rural development effort to reach a
target population of small farmers who were
traditionally neglected by the agricultural
research system. The initial elements used were
on-farm experiments rather than formal sys-
tems approaches, and most projects were imple-
mented by agricultural researchers trained in
traditional disciplinary methods with no knowl-
edge of systems theory. FSR began in a range of
research and rural development projects within
existing institutions, and these, like the man-
power, were organized to operate the more tra-
ditional approach. These institutions pursued
the conventional process; field research activi-
ties were identified and designed by senior
researchers working on an experiment station.
Extension agents provided small farmers with
technical recommendations based on nation-
wide research results which, in the best sce-
nario, included a trial from the local region.
It can rightly be claimed that FSR in LAC is
characterized by heterogeneity in its applica-
tion, its methodology and its institutional con-
CAB International 2000. A History of Farming
Systems Research (ed. M. Collinson)

text. The initial introductions of FSR were con-
centrated in the Puebla project in Mexico, the
Caqueza project in Colombia and at the Tropical
Agricultural Center for Research and Training
(CATIE) in Guatemala. I was fortunate enough
to have personal experience in each of these
three: I was a young professional in the forma-
tion of the Caqueza team and went to the
Puebla project for in-service training. Later, fol-
lowing the Caqueza project and graduate work,
I moved to CATIE which was assembling a
strong team of systems researchers. This paper
is a personal interpretation of the initial devel-
opment of an FSR approach in Latin America
and it is a difficult task for a practitioner used to
writing up experiences in the most formal way
possible; a defensive strategy adopted to pre-
empt the permanent criticism from biological
researchers of the lack of 'science' in FSR appli-
The political context of development in Latin
American countries played a strong role in how
FSR was introduced. Neglect of the peasant sec-
tor brought an emphasis on the use of FSR
methods in the field by organizations with a
social conscience. As a young professional 22
years old, I found the opportunity to share my
theoretical training and political ideas with very
small farmers living in poor conditions very
stimulating. In those early days of the 1970s,

Chapter 2

FSR relied on readily available methods and
there was limited interest in conceptualization
or the evolution of theory. Some description of
the political context is important to understand
the fragmented and ad hoc emergence of FSR,
followed by an account of some of the experi-
ences in the three projects which made an
impression on me. Finally, some comment on
the evolution of the FSR concept and its
methodology is offered. This is perhaps the most
difficult part, since the rapid evolution of the
concept often contradicts the interpretation a
young technician made some 25 years ago.

2.1.2 The context of agricultural
research in LAC
Although FSR became a phenomenon in the
early 1970s, changes in the LAC agricultural
sector during the 1960s influenced its applica-
tion. The national agricultural research insti-
tutes (NARI) established in the 1960s
internalized a strong influence from the USA
'land grant system' through technicians trained
in the USA and in research cooperative pro-
grammes supported by the USA in Mexico and
Colombia in the 1950s1. Agricultural research
was disciplinary and crop-oriented, and its
institutions were designed to operate through
research stations. From the beginning, priorities
were set from the supply side, mainly in
response to the import substitution policies of
governments and the scientific interest of
researchers. The publicized success of the Green
Revolution supported by the International
Agricultural Research Centres (IARCs) rein-
forced this traditional approach.
In the 1960s the contribution of agricultural
products to the gross national product (GNP)
and the increasing importance of the multilat-
eral agencies made agricultural research an
important area for national policy. These policies
were driven by the need to produce cheap food
for the urban and industrial sectors and support
services such as agricultural credit, marketing
and some physical infrastructure, were strength-
ened to complement agricultural research and
extension. During the 1960s land reform and
agricultural development projects, implemented
in most LAC countries, highlighted the gap
between USA agriculture and LAC, on the one
hand, and the traditional small-farmer subsector

and the commercial subsector on the other.
Strengthening the extension services and the
NARIs with the collaboration of the interna-
tional community was not sufficient to close
these gaps. NARIs did reinforce their research on
rural issues that were known to influence pro-
duction and technical change among farmers.
Programmes such as agricultural economics,
communications, adult education, social organi-
zation and home economics were incorporated
into their structure. This required the skills of
social scientists and other growing professions
that could contribute new information about
farm organization, farm production strategies,
land tenure and related issues, farmers' organi-
zations, their priorities and values, their commu-
nication systems and their off-farm activities.
This new information improved the analytical
capacity of NARIs, while better definition of the
target population improved their responses.
Both land reform and agricultural develop-
ment projects put extensionists and some
researchers in the field with farmers. The orga-
nization of these projects brought together, per-
haps for the first time in LAC, most of the
available agricultural support services as well as
technical teams from different institutions and
disciplines. Though it could not be hailed as a
coordinated effort, technicians in this working
environment vigorously identified with the
small producers, land reform beneficiaries and
settlers. With hindsight it was probably a key
step in transforming research to reach small
farmers. At the same time, new information on
farmers' production strategies, technical prac-
tices, the social organization of production and
local markets brought new evidence of the pro-
ductivity gap and the missing link between agri-
cultural research results and the capacity of
this traditional subsector to assimilate and uti-
lize technical information.
The influence of multilateral agencies was
more evident on the design and implementation
of the Integrated Rural Development (IRD) pro-
jects initiated at the end of the 1960s. These
were specifically aimed at improving production
and living conditions of small traditional farmers
through agencies and policies designed to help
the consolidation of the industrial sector. In a
number of countries these IRDs were important
components of national development policy, and
in some cases, as in Brazil, Colombia, Ecuador,

FSR: Origins and Perspectives 15

Honduras and Mexico, this lasted until the mid
1980s. Most IRD schemes posted interdiscipli-
nary technical teams in the field as well as sup-
port institutions. Although rural development
components other than technical change were
considered important, increasing agricultural
production and productivity usually became the
central focus, supported by credit programmes
which were redesigned for this purpose.
Implementation increased the evidence of the
limited use of the existing technology in the pro-
duction conditions of small farmers.
These donor-supported projects provided a
context which allowed a number of technicians
to question the capacity of research institutions
to generate technical recommendations for this
target population. Conceptually, the rationality
and efficiency of small farmers were already
demonstrated2. Pragmatically, the rejection of
technical recommendations by small producers
was a daily reality for the field teams. Some
began to test different arrangements of known
technological packages at the farm level, some-
times maintaining practices and components
already used by producers3.
The pioneer work in the late 1960s and the
early 1970s in Mexico, Colombia and Central
America was pragmatic; standard research
designs, farm trials and biophysical analyses
were utilized. In response to national policies,
staple foods (grains) were preferred to high rev-
enue crops. Farmers' crops were taken as the
basis for designing technological alternatives
and the emphasis was on providing support ser-
vices, particularly agricultural credit. Field
teams were formed in most cases by young tech-
nicians with no experience. Expatriate technical
advisors were often involved and the rigid insti-
tutional organization was not changed.

2.1.3 Initial FSR applications
Puebla, Mexico
Most LAC practitioners recognize the Puebla
Project in Mexico created in 1967 as the first
agricultural research activity closely related to
farmers' production patterns. Although the
examples of Borgo-a-Mozano, the Comilla
Project and the Intensive Agricultural Districts
Program in India were known, Puebla intro-
duced the on-farm adaptation of production
technology developed at experiment stations

which considered farmers' cultural and capital
limitations to adoption. The project was
designed and technically supported by CIM-
MYT, the International Maize and Wheat
Improvement Center, in partnership with the
Chapingo Graduate School.

Caqueza, Colombia
In 1970, the Colombian Agricultural Research
Institute (ICA) initiated the Caqueza project
with the collaboration of the International
Development Research Center of Canada
(IDRC)4. The field team of the project was
formed by young technicians: agronomists, a
veterinarian, a sociologist, a home economist,
two advisors from IDRC and a number of field
assistants trained at the intermediate level.
Activities related to on-farm research, exten-
sion, planning and evaluation were established
from the beginning. Some researchers from
experiment stations, graduate and undergradu-
ate students, foreign volunteers and some con-
sultants carried out research activities in the
project, producing over 100 publications in the
first 6 years.
Recruited as a team member, I was
instructed to go to the field and 'get the region
developed'. This missionary approach reflected
two key problems. First, senior managers had
no clear idea of the type of project they wanted.
Second, the integration of traditional
researchers in the project's field activities was
not systematically organized and became a
major difficulty. Several efforts were made by
senior project managers to plan joint research
in the field, but these were weakly implemented
throughout the first 10 years of the project. The
plans for senior researchers to assist the techni-
cal team were never realized, and indeed these
researchers became critical of the validity of
on-farm research. I was sent for in-service
training to the Puebla Project in Mexico which
was very effective in providing training for a
number of young technicians from different
countries. The Puebla approach and its compo-
nents were extrapolated to many other projects,
including on-farm technology trials and the
analytical research orientation through stu-
dents and university faculty members. The
training instilled a much stronger conceptual
framework for subsequent activities in on-farm
research in Caqueza.

16 Chapter 2

Institutionally, in Caqueza, the major preoc-
cupation was to provide the capacity to adapt
technology tested in experiment stations to
farmer's circumstances and to obtain high
adoption rates among direct beneficiaries.
Despite the great efforts made by some mem-
bers of the team to develop a framework for the
entire project, the farming systems concept was
never explicit. However, terms such as multiple
cropping, on-farm research, multidisciplinary
teams, farmers' constraints, farm types, farm
production components and the understanding
of the farm as a decision unit were all used in
the implementation of the project. Most Latin
American practitioners at that time wrote for
field colleagues, but their writings were not
reviewed by their peers, formally published or
translated into other languages. Conceptual
contributions to FSR began to be made only
after some years of field experience5. In the mid
1970s Richard Harwood's writing on his expe-
rience in Asia, and David Norman's on his time
in Africa, offered useful frameworks which
were, in turn, complemented by learning from
other applications in Latin America6.
The concepts of team and institution, the
institutional challenges and the strong identi-
fication with small farmers were, in my view,
the key motivations for the small group of
young technicians that began their profes-
sional activities in the Caqueza Project. The
level of involvement of every one was remark-
able, substituting for the conceptual elabora-
tion and development theory unavailable to
them. The possibility of constructing a differ-
ent institutional model, the open-ended oppor-
tunity to learn about small farmers, the search
for solutions to real problems, alternative insti-
tutional instruments and, above all, applied
research methods, were powerful reasons to
maintain the professional interest and the
motivation required to initiate a long explo-
ration of the application of FSR to agricultural
It has to be said that the socioeconomic
dimension still lacked methodological integrity.
More than a year was needed to complete a
regional diagnostic that made little contribu-
tion to the understanding of farmers' produc-
tion strategies. Indeed, a number of on-farm
experiments were concluded and an extension
strategy was in place before the results of diag-

nosis were available to the technical team.
Technical integration between biology and
social technicians was a difficult process. A
common view of the project was shared by all
team members but different working methods
in the field caused important discrepancies.
These differences frequently discriminated
against social research and, in some cases,
resulted in personal conflicts.
One important result of the Caqueza Project
was the recognition of the need to adapt institu-
tions to the circumstances of small farmers. The
project generated the so-called 'buffer institu-
tions' as pilot institutional programmes to facili-
tate adoption of recommended technology, and
some changes to institutional services were
introduced when the IRD projects and, some
years later, the national fund to support rural
development activities, were put in place.
While I attended graduate school in the mid
19 70s I became better acquainted with the liter-
ature. There were still very few publications clas-
sified under FSR but it was a great opportunity
to make sense of theory, working approaches
and farmers' reality. It confirmed the need to
develop methods and institutional arrange-
ments to respond to farmers' real conditions.

CATIE, Costa Rica
In the early 1970s, CATIE initiated on-farm
research programmes that evolved into major
applications of the systems approach with the
support of the United States Agency for
International Development (USAID) and largely
through the influence of Richard Bradfield.
These efforts lasted more than 10 years and
influenced most research programmes at CATIE
and a number of institutions, researchers and
graduate students in Central America.
My work with the team at CATIE was the
third major influence in drawing me into advo-
cacy for FSR. The CATIE technical team was well
qualified and equipped. It was formed by well-
trained agronomists, animal scientists, agricul-
tural economists and ecologists with a support
team for biostatistical data processing. A num-
ber of papers, training materials and MSc theses
were published both internally and externally. In
comparative terms, CATIE publications com-
bined empirical experiences with conceptual
development and constituted the state of the art
on the application of FSR in Latin America in

FSR: Origins and Perspectives

the early 1980s. A peculiarity of the early CATIE
FSR team was the decision to work indepen-
dently from the ideas developed at Puebla and in
Guatemala. Strong personal attitudes dictated
the approach developed. It did not incorporate
the Guatemalan 'sondeo' approach, or the ideas
of farmer organization and the provision of sup-
port services. Since CATIE is mandated for
research and training only, extension, technol-
ogy transfer and institutional adaptation were
virtually absent from the CATIE work schedule.
The first exchange of concepts and some experi-
ence from El Salvador led to the design of a big
central experiment, testing numerous interac-
tions and different topological arrangements, at
Turrialba station. This generated a number of
novel ideas among the research team but was so
complex that analysis of the experiment was
never completed!
After a team member visited the International
Rice Research Institute (IRRI), the conceptual
framework being used there by Harwood and
Zandstra was introduced into a regional project
that CATIE developed in Central America. Basic
system concepts were applied to agricultural
development, allowing the construction of a
systematic body for teaching and training pur-
poses7. This conceptual base brought about the
development of different methodological phases
for applying the systems approach on the field:
area selection, diagnosis or characterization,
technological alternatives, design, on-farm
research, validation and dissemination, includ-
ing a feedback mechanism to generate results.
In every case field work was initiated as an
empirical test of a conceptual elaboration8. The
improvement of production systems based on
farmers' practices captured most design efforts,
the programme included better understanding
of the biophysical relations among plant, soil,
water and crop management. Later on, agro-
forestry practices for resource conservation and
rationalization were developed and added as
technological alternatives for improving farm-
ers' production systems.
Economic and social analysis became an inte-
gral part of the CATIE approach. The economic
analysis of agronomic trials introduced by CIM-
MYT was complemented with a range of analyti-
cal tools, putting emphases on two aspects:

* Farmers' constraints due to both farm and
off-farm circumstances.

* The impact on farm activities as a whole and
the adoption possibilities.

These analyses were useful to understand the
production system as a decision unit. The appli-
cation of discrete and linear programming
models to agroecosystem analyses helped tech-
nicians and students understand the basic eco-
nomic relationships involved in farming

2.1.4 Early evolution of the FSR approach

Since CATIE has a regional mandate, a number
of institutions in almost every country were
involved in the application of FSR as well as in
the development of methodology. A consider-
able number of technicians were trained on
those concepts and their application, together
with students from South America. Projects
following the systems approach were also
implemented in a number of countries, influ-
enced by either these pioneer examples or by
donors that adopted the approach for the devel-
opment of small farming areas in Latin
America. Such projects were found for example
in Ecuador, Honduras, El Salvador, Bolivia,
Peru and Brazil. Thus the initiation of the FSR
application in LAC was extremely dynamic and
volatile. Changes in the development process,
multiple applications to different conditions,
the involvement of the scientific community
and a vigorous exchange of information con-
tributed to rapid evolution in the first 10 to 15
years. These early changes can be summarized
in four points:

* Wider options to be considered in the design
of technological alternatives for a farming
* The adaptation of some research and exten-
sion programmes to the FSR approach and
the movement of a number of researchers
off the experiment station.
* The introduction of socioeconomic analyses
to consider the complete farm as the focus
within a hierarchy of systems.
* The evolution of the idea that recommenda-
tions might include non-agronomic inter-
ventions based on comparative advantage,
commercial strategies or farm management

18 Chapter 2

In the early 1980s, several attempts to institu-
tionalize the FSR approach took place all over
the continent. NARIs in El Salvador, Panama,
Guatemala, Colombia, Ecuador, Peru and
Brazil, among others, formally announced their
decision to adopt this research approach.
Nonetheless, research institutions have main-
tained, in general terms, a centralized organiza-
tion that favours the traditional research
approach. On the other hand, it has to be recog-

nized that the concept of FSR has been accepted
for most agricultural researchers and rural
development agents, who have used some of its
principles to design their research work.
Personally, the 5 years at CATIE consolidated
my opinion: by the time I joined IDRC as a pro-
gramme officer I was a firm advocate for FSR.
convinced that a systems approach is a power-
ful tool to analyse agricultural development at
any level.

1. Moreno, R.A., 1984. Estado actual de la investigation en sistemas de finca en America Latina y El Caribe.
In: Compton, P. & DeWalt, B. (Eds) El sorgo en sistemas de production en America Latina. CIMMYT-
2. Schultz, T.W., 1964. Transforming Traditional Agriculture. Yale Press.
3. Moreno, R.A., 1984. (Op. cit.).
4. Zandstra, H., K. Swanberg, C. Zulbert & B. Nestel, 1979. Caqueza: Experiencias en desarrollo rural.
Centro Internacional de Investigaciones para el Desarrollo, Bogota.
5. Escobar, G. & K. Swanberg, 1972. Metodologia para investigation del nivel de vida como component
de una estrategia de desarollo. ICA. Reg. 1. Proyecto de Desarrollo Rural del Oriente de Cundinamarca.
Mimeo, Bogota.
Myren, D.T., 1971. The Puebla Project: a Development Strategy for Low Income Farmers. Seminario
sobre Estrategias de Desarrollo del Pequeo Agricultor. Ohio State University, Ohio.
Institute Colombiano Agropecuario, 1972. Justificacion de la estrategia y un modelo para el desar-
rollo rural. ICA. Reg. 1. Proyecto de Desarrollo Rural del Oriente de Cundinamarca. Mimeo, Bogota.
6. Hildebrand, P.E. & E.C. French, 1974. Un sistema salvadore de multicultivos. Su potential y sus
problems. In: Conferencia Sobre Sistemas de Producci Agricola Para el Tropico. CATIE, Turrialba, Costa
Zulberti, C.A., 1974. Rural Development Information Requirements. IDRC: Simposio del Personal de
Campo de Ciencias de Alimentos y Nutricion. Ottawa, Canada.
7. Hart, R., 1981. Agroecosystems. CATIE, Turrialba, Costa Rica.
8. Escobar, G. & R. Moreno, 1984. Desarrollo de Tecnologa de Producci Agricola: Enfoque Metodologico y
Aplicaci Empirica. Trabajo presentado en el Taller Internacional de Sistemas Agricolas. FAO. CATIE,
Santiago de Chile.

Peter E. Hildebrand

During the 'Sondeo', or reconnaissance for the survey, we found that most of the farmers were on the steep.
rocky hillsides, not in the valleys where the experiment station was and where the first 'on-farm' trials were
being conducted.

2.2.1 Early days

For me, FSR-E started with multidisciplinary
research as soon as I started my professional
career as an agricultural economist. At the
Range Management Department of Texas A &
M University back in 1959, I worked very

closely with plant and other biological scien-
tists. Moving to Colorado State University in
1961, I was part of a joint project with the
Agronomy Department, alongside William
Schmehl of the Shaner et al. 1982 publication
on Farming Systems'. This groundwork was
reinforced by my first overseas assignment from

FSR: Origins and Perspectives

1964 to 1967 with an engineering firm, work-
ing for the West Pakistan Water and Power
Development Authority in Lahore. As Chief
Economist and then Chief of Planning, it
became evident to me that a coordinated, multi-
disciplinary effort, with a common objective, in
this case irrigation reclamation project design,
was a powerful approach to R & D.
Towards the end of my third overseas post,
assigned by the University of Nebraska to the
ICA in Colombia from 1968 to 1972, I began to
work with small, very limited-resource farmers
in an on-farm research and demonstration pro-
ject2. It convinced me that small farmers are
neither backward nor against change, but are
very willing to adopt new technologies when
these are appropriate to their conditions.
These experiences were good preparation for
my assignment by the University of Florida to
work with Centro National de Tecnologia
Agropecuaria (CENTA) in El Salvador in 1972
on the design and testing of technology
expressly for the many small, limited resource
farmers of that country. Here, the influence of
Richard Bradfield's multiple cropping system3
was strong and was partly responsible for the
establishment of the multidisciplinary Multiple
Cropping (Multicultivos) Program in the
Agricultural Economics Department of CENTA
which I headed. One objective of the programme
was to design technology for scarce land and
abundant labour conditions. National policy was
to reduce the level of vegetable imports, mostly
from Guatemala. So the programme emphasized
increased vegetable production without reduc-
ing the output of maize and beans, the main sta-
ple crops in the country. Multicultivos moved
into on-farm trials in its second year, 1973,
becoming a national programme in 1976.
The influence of Agricultural Economics on
Multicultivos was due to two factors. First, the
system was developed by the Department of
Agricultural Economics. Second, CENTA recog-
nized that the combination of crops into sys-
tems to improve resource use was an area of
farm management. It was logical, therefore, to
base Multicultivos in the Department4.

2.2.2 ICTA, Guatemala
At the same time, the Institute of Agricultural
Science and Technology (ICTA) in Guatemala

was being created as a semi-autonomous agri-
cultural research institute. It was charged with
developing technologies for basic grains grown
by the small and medium-sized farmers who
produced two-thirds of the country's require-
ments, but who had been by-passed by previ-
ous technology development efforts5. In May
1974, I met with Directors of ICTA and Joe
Black, Director of Social Sciences of the
Rockefeller Foundation, to discuss the responsi-
bilities of a position of Coordinator of Rural
Socio-economics (SER).
They explained the situation. The
Guatemalans felt that if the small farmers of
the country were to be included in economic
development, it was essential to have a better
understanding of their needs and limitations -
something that the social sciences should be
able to provide. The aim was an institute in
which social sciences were integrated with bio-
logical sciences to ensure that research was, in
fact, oriented towards the needs of the small
farmers. There was general agreement at that
meeting on the possible scope of SER's work,
but there was virtually no discussion of the
methodology to be used6.
In 1974, the Rockefeller Foundation
appointed me the Coordinator of Rural Socio-
economics of ICTA and in January 1975, SER
presented a seminar to the Institute outlining
our role as we perceived it, and our plans for the
year7. We planned to help the commodity pro-
grammes understand the needs of small and
medium farmers through surveys and by help-
ing these programmes design and analyse their
trials. SER had no resources to conduct trials on
its own. Our contribution was to influence the
substance of the trials being undertaken and
was important for ICTA. Early results, however,
were not without conflict.
One of our first projects was a survey in the
eastern part of Guatemala. During the 'Sondeo'
or reconnaissance for the survey, we found that
most of the farmers were on the steep, rocky
hillsides, not in the valleys where the experi-
ment station was and where the first 'on-farm
trials' were being conducted. In collaboration
with the bean programme, we proposed trials
on the poorest (and only sloping) part of the
new station, using bullocks to plough rather
than tractors to keep the conditions as close as
possible to those of our small-farmer clients.

Chapter 2

The Regional Director, however, wanted to
homogenize the station. Our planned use of
bullocks, and our plans to do without fertilizer,
did not fit with this concept, so in March 1975,
we decided, in consultation with the Technical
Director and the Regional Director to move off
the station and conduct our farm trials on
rented land. It was the best thing that could
have happened8.
The result was the first on-farm trial to be con-
ducted under the real, very difficult rocky hillside
conditions of the farmers in that part of the coun-
try. Although the land was rented, local farmers
were very much involved in the design of the trial
and resulting technology. The site was visited by
an impressive list of individuals, who, along with
the Institute, learned a great deal about the
nature of 'on-farm trials' from the experience.
This showed that it was necessary to conduct tri-
als under these conditions if we were really to
understand how and why the farmers did what
they did. Eventually it was recognized that SER
should have a budget to conduct its own on-farm
trials. The difficulties of working under the trying
conditions of the hillsides also forced the Institute
to get to grips with the reality of working for the
small-scale farmer.
In January 1976, SER conducted an agroso-
cioeconomic survey around Tecpan in the cen-
tral highlands, with the help of one person from
the bean commodity programme and one from
the local 'Technology Testing Team' as a foun-
dation for on-farm trials that year. The trials
were again conducted on rented land, but with
ample input from local farmers. Three classes of
farmers had been identified in the survey and
treatments were designed for each. In the sec-
ond year SER initiated 'Farmers' Tests' with the
technology generated in the trial.
Our experience in Tecpan demonstrated the
benefits of a multidisciplinary team gathering
information based on local survey results before
the trials began. And it helped us understand
the value of using local farmers both as advisers
and as sources of labour in the field trials.
Tecpan showed the value of simple technology,
based on that already found in an area, with
only minimal changes9.

2.2.3 The Sondeo
By January 1977 we were beginning to doubt
the need for the full-scale surveys being con-

ducted by SER in conjunction with the com-
modity and production teams. Because they
were participating in the field surveys, the field
teams were analysing and interpreting during
the survey. The purpose of the surveys, to pro-
vide information for the regional teams to use in
orienting and planning their work. was accom-
plished by the time the field work had been com-
pleted. Little was gained by a written report
delivered several months later. We were finding
that the first impressions gained during the
'Sondeo' or preliminary survey, to obtain an ini-
tial understanding of the area and design an
appropriate questionnaire, were correct.
Although we continued to do the full surveys
during 1977, we conducted 'Sondeos' only in
two areas. We found the Sondeo provided a great
deal of useful information for planning on farm
experiments in both instances.
We conducted just one survey and three
Sondeos in 1978 and, for the first time, a Sondeo
in Moyuta in south-eastern Guatemala was
conducted without any plan to follow it with a
survey. We found that much more information
was gained than if a Sondeo was carried out as a
prelude to a survey like that done in Jalapa in
the Eastern Highlands that same year. In
September 1978 in Zacapa, we began to firm up
the Sondeo methodology.
While ICTA was primarily a crop research
institute, it did have a small livestock compo-
nent. Our first livestock survey in early 1978
was problematic. For the first time we were
unable to understand the farming system being
studied. By concentrating on the livestock and
not considering either the crops, or the interac-
tion of the two, the livestock practices did not
make sense. A follow-up survey including both
was successful. After that we began looking at
both crops and livestock in every Sondeo,
regardless of whether one or the other was our
primary focus. By 1979, the 5-week long
Sondeos, conducted early in the year, had
become the accepted method for obtaining pre-
liminary information for an area.
In addition to increasing confidence in the
results of Sondeos, another reason to eliminate
questionnaire-based surveys was the farm
record project, begun in 1975 as an additional
method of obtaining crop production informa-
tion in areas where 'Technology Testing Teams'
were assigned'1. The project was conceived as a

FSR: Origins and Perspectives

crop, not a farm record programme. Data were
kept on individual crops or crop associations,
not on the farm as a whole. The crop record
project grew to a full national project that has
continued over the years. Between 1975 and
1978 it expanded from one area to 11, from
three crops or crop associations to 34, from 40
individual records to 583 and from 390 ha to
over 140011.

2.2.4 Index of acceptability

Perhaps one of the most important features of
the organization and management of research
at ICTA was charging the socioeconomic unit
with the responsibility for the technical evalua-
tion of its research.
As ICTAs Five Year Plan of 1975 said:
Technical evaluation of ICTA will be in the charge
of SER. The reason for putting this group in
charge of evaluation is to assure an orientation
not only of the agronomic factors, but also of the
socio-economic factors of the farmers. By doing
this, the institute hopes to have an orientation
directed towards resolving the problems of the
small and medium farmers of the country and
avoid investing in projects that would have little
potential for increasing the income of the clients
or increasing national production. ... The evalua-
tion process will begin with the development of
new projects, continue during the execution to
assure that it is being done under conditions rele-
vant to the farmers, include the evaluation of rec-
ommendations and of the results of the
technology when it is placed in the farmers' hands
by determining the acceptability of the technology
and finally will close the circle with recommenda-
tions based on an analysis of the previously
described process. 12

To help meet this responsibility, SER developed
an Index of Acceptability which is still in use
today13. This index, which measures accepta-
bility not acceptance of a technology,
indicates the extent to which farmers were con-
vinced to continue and widen the testing of a
technology tried in on-farm tests the previous
season. As finally developed, the index calcu-
lates the percentage of farmers who tested a
technology the previous year, multiplied by the
percentage of area of that crop on which they
are using the technology, divided by 100. In
1976/77 the index for maize in one project area
clearly showed that:

* Technologies with a number of components
were essentially never acceptable in total.
* Some components such as soil insect control
and planting date were not acceptable.
* Improved seed and planting distance (mech-
anized) were acceptable.
As the number of components included in trials
dropped from eight to four, the average
Acceptability Index increased from 19.8 to
47.6. It is interesting that fertilizer use,
included in the trials and recommended for
farmers, was clearly not acceptable, a finding
confirmed from farm records14.

2.2.5 University of Florida, Gainesville
Ken McDermott, the project manager for the
USAID-funded Shaner et al. report, made occa-
sional visits to ICTA in Guatemala. On one of
these I was showing him around the country
and explaining the differences in farming sys-
tems as we drove. He said, 'Pete, I can see that
you understand the differences in these systems
that to me look very much alike. Can you train
others to do the same thing?' This question and
challenge stayed with me and while on sabbati-
cal leave from the Rockefeller Foundation at the
University of Florida (UF) in 1980, I decided to
accept the challenge. Rather than return to the
Foundation, I chose to stay on at UF and
became the Coordinator (a non-title) of the
Farming Systems Program (a non-entity) in
IFAS (the Institute of Food and Agricultural
Sciences) at UF
The programme had a number of facets. We
created a short course training programme
which became part of the USAID-funded
Farming Systems Support Project (FSSP) head-
quartered at UF. When the FSSP terminated, the
short course moved into the International
Training Division (ITD) which I directed for 3
years. We taught the first Farming Systems
Research-Extension Methods course at the
graduate level in spring semester 1980, and it
has been taught every year since. With funding
from the Vice President for Agricultural Affairs,
we initiated Farming Systems Assistantships
which brought many fine students to the gradu-
ate programmes in agronomy, agricultural eco-
nomics, soil science, forestry, and agricultural
education and communication. Funding from

22 Chapter 2

the US Department of Agriculture and IFAS
helped us create the North Florida Farming
Systems Project in a six county area. Farming
systems is now the largest minor for graduate
students in the College of Agriculture, while a
concentration (major) in farming systems at the
MS degree level in the Agricultural Education
and Communication Department has a number
of graduates and current students.
Even though the FSSP ceased to exist in
1987 and the North Florida Farming Systems

Program shortly after that, the graduate pro-
gramme in farming systems is still going strong
and the short course programme is still run-
ning. After nearly 20 years, the Farming
Systems Program is flourishing at UF and farm-
ing systems methodology remains an important
tool for many of the graduate students cur-
rently involved in conservation and sustainable
development curricula in many departments
and colleges on campus.

1. Shaner, W.W., P.F. Philipp & W.R. Schmehl, 1982. Farming systems research and development: guide-
lines for developing countries. Westview Press, Boulder, CO.
2. Andrew, C.O. & P.E. Hildebrand, 1993. Applied agricultural research: foundations and methodology.
Westview Press, Boulder, CO.
3. Bradfield, R., 1966. Toward more and better food for the Filipino people and more income for her farm-
ers. ADC Paper, the Agricultural Development Council, New York.
4. Hildebrand, P.E. & E.C. French, 1974. Un sistema salvadore de multicultivos. Su potential y sus
problems. In: Conferencia Sobre Sistemas de Producci Agricola Para el Tropico. CATIE. Turrialba, Costa
5. Waugh, R.K., 1975. Four years of history. Institute de Ciencia y Tecnologia Agricolas (ICTA). Guatemala.
6. Hildebrand, RE., 1979. Incorporating the social sciences into agricultural research: the formation of a
national farm systems research institute. Institute de Ciencia y Tecnologia Agricolas. Guatemala and The
Rockefeller Foundation, New York.
7. Hildebrand, P.E., 1975. El papel de socioeconomia rural in el Instituto de Ciencia y Tecnologia Agricolas
(ICTA). ICTA, Guatemala.
8. Hildebrand, P.E., 1979. (Op. cit.).
9. (Ibid.).
10. Shaner, W.W., P.RE Philipp & W.R. Schmehl, 1982. (Op. cit.).
11. Hildebrand, RE., 1979. (Op. cit.).
12. (Ibid.).
13. Hildebrand, P.E. & E Poey, 1985. On-farm agronomic trials in farming systems research and extension.
Lynne Rienner Publishers, Boulder, CO.
14. Hildebrand, P.E., 1979. (Op. cit.).

Richard Harwood

The major resource allocation problem which faced systems research in the 1970s. as it does now. is that
systems understanding is one step removed from actual impact, and is not as obvious in its contributions to
eventual improvement.

The framework for Asian FSR and extension in
the 1970s can be traced directly from the work
of Dr Richard Bradfield. His research took place
within the context of a centuries-old evolution

of intensive cropping systems by Asian farmers
and was influenced along the way by those
researching farming systems, particularly in
Africa. Such authors as Nye in 19611 on
organic matter and nutrient cycles under moist
tropical forest, Papadakis in 19652 on crop

FSR: Origins and Perspectives

ecology and Ruthenberg in 19713 on farming
systems integration, not only influenced
Bradfield, but had a major influence on my own
thinking. FSR in Asia became increasingly col-
laborative with Asian scientists in several coun-
tries under Bradfield in the late 1960s and then
through the IRRI-based Asian Network pro-
gramme through the mid 1970s. Underlying
assumptions, methodologies and key areas of
focus evolved as participation broadened. This
summary captures some of that evolution.
Bradfield wrote in 1964 that his first contact
with rice growing in Asia dated to the arrival of

Phase of


No farmer involvement necessary

Felix Ponnamperuma on the Cornell campus
from Ceylon in the early 1950s. Shortly after, in
1955-56, Robert Chandler and Bradfield trav-
elled throughout Asia as representatives of the
Rockefeller Foundation to make recommenda-
tions on how to improve agricultural produc-
tion across the region. In his 1956 final report
to the Rockefeller Foundation, Bradfield said
that 'rice yields in all countries of the area vis-
ited, with the possible exception of Japan and
Taiwan, could be at least doubled by further
research and education and that in many of the
tropical areas with adequate water supply, food


- - - -- -
Of technical and support factors Of s
Physical Biological Support systems I Resour

Water availability Varieties Extension Flows a
observation Temperature Diseases Credit (la
and Sunshine Insects Input distribution capital
ascription Soil Weed control Price
n produc- Government policy I
on areas)

I \

Y-- -- -- -- -- -- -- -- -- -- 4- -
technology. Feedb
for a

\ I
.. -\

Feedback on \
changes needed -
Test A. Test for physical an
biological adaptation


mer involvement necessary

socio-economic factors
ce quantities Social attitudes

nd relationships Tr
nd, labour,
, management)




'ack on needs
ttainment of


Feedback on
changes needed
id /

B. Test for compatibility
with farming systems


Fig. 2.3.1. A framework for cropping systems research and extension (from IRRI, 1975, p. 11).

24 Chapter 2

production could be greatly diversified and
again doubled in quantity by more intensive use
(of other crops) during the dry season'4. This
report recommended the establishment of IRRI.

2.3.2 Bradfield at IRRI
In support of his rationale for increased yield
potential, Bradfield began to compare water
availability, solar energy and temperatures
between Ithaca, NY, his research base for sev-
eral decades, and Los Bafios in the Philippines.
His research showed that resources in tropical
Asia were grossly under-utilized. By 1964 the
Bradfield programme at IRRI was based on the
following assumptions:
* Family farms will continue to predominate
in Asia, but will increase in size to 4 or 5 ha.
* Rice will continue as the basic staple.
* Water availability will limit non-rainy sea-
son rice production.
* Water can be developed for year-round use
in most of Asia.
* Average rice yields can be increased 2-5
times over the next two human generations.

* Farmers' need for income will require a dou-
bling of farm size, the yield of rice increased
2-4 times, and non-rice crops grown in dry
* Multiple cropping intensity must be doubled.
* Field operations must be minimized and
From these assumptions, Bradfield developed his
system which, by 1971. he termed the 'maximum
cropping project'5, making maximum use of geo-
physical resources, crop rotations and the integra-
tion of mechanization6. Bradfield had modified
the Taiwanese ridge-furrow system to fit a very
versatile hand tractor, the Landmaster7. The
Taiwanese system captured in the colour slides
that he gave me on my arrival at IRRI in 1972,
bore a remarkable resemblance to his mechanized
'IRRI' system, but used hand labour and animal
traction. The Landmaster could successfully do all
of the soil shaping and tillage needed for the
intensive production system shown in Table
2.3.1. In his most intensive cereal system, shown
in Fig. 2.3.1, the rice-sorghum system with
ratoon crops produced at its maximum 30 t of
grain ha-1 in a 12-month period in 1968.

Table 2.3.1. Typical maximum cropping pattern (from Bradfield, 1970, p. 239).
Mean yield Mt ha-1 Gross
Date of income
No. By- Price ha-'
Crop Planting Harvest of days Tillage operations Crop products* (US$) (US$)

Rice 1 June 10 Sept. 102 5 5 100 500

potato 15 Sept. 4 Dec. 100 25 20 10 1000

Soybeans A
(dry) 27 Dec. 17 Mar. 85 2.5 100 250

Maize 1 Mar. 5 May 66 e/ ..__ 1 40,000
(ears) 15 0.02 800

(vegetable) 1 May 1 July 60 J J 6 6 100 600
(green pods)

Total 3150

* By-products: straw, vines and stalks.

FSR: Origins and Perspectives

Apparently this yield was never repeated.
Twenty-five tons was more commonly achieved
both by Bradfield and by those of us who fol-
lowed. Bradfield also organized training classes
for field researchers from Indonesia and Ceylon
in the late 1960s. Prompted by his visits and
training, the Maha Illiphalama site in Ceylon
became the first national training centre of the
1970s era in Asia.
Bradfield should be credited for three major
* He proved that the yield potential of cereal
and feedgrain, on an annual basis in the
tropics, could equal or surpass that of tem-
perate areas, dispelling the myth that tropi-
cal areas had low production potential.
* He designed an integrated system of soil,
water and crop management around a very
particular kind of power source, all intended
to maximize use of soil, water, temperature
and light. The system was designed for a
farm of 3 to 4 ha in size, having full access to
inputs and markets. The trials were all done
on field-scale plots of 0.25 ha. To my knowl-
edge this was the first instance of such com-
plete system research design and test to
utilize a prescribed resource base.
* He recognized that specialized systems train-
ing was necessary for workers in integrated
systems which were designed around
resource-use concepts.
Bradfield was wrong, however, in several of his
assumptions. He did not foresee the rapid indus-
trialization and labour shortages, even with
mechanization, which have actually decreased
cropping intensity in several Asian countries,
most notably Taiwan. His assumption of
increasing farm size was correct in a way, again
as in Taiwan with structural transformation
the numbers engaged in agriculture fell and
average farm size increased.

2.3.3 Making progress
When I followed on from Bradfield at IRRI in
1972, I became convinced that his system
design capability, learned over a life time, could
be conceptualized and taught. The need for that
capability and for the excellence in technologies
required by integrated, high productivity sys-
tems, guided the IRRI programme of the 1970s.

Working with Gordon Banta, an agricultural
economist, IRRI's work for the next 4 years con-
sisted of:
* Mastering the Bradfield maximum produc-
tion methods, quantifying results, and
reviewing underlying assumptions (during
* Increasing linkages to collaborators in Asian
national programmes the Asian Cropping
Systems Working Group was formed.
* Developing a process for cropping systems
* Developing component technologies for
intensive systems.
* Assembling an expanded research team to
deal with the multidisciplinary demands of
Asian systems.
* Selecting multiple sites and methodology
development for systems research.
* Training for outreach site coordinators.
The first four activities were underway in 1972.
By 1973 a conceptual framework was taking
shape and training had started. By 1974 multi-
location testing occupied most of our time, with
methodologies for system research in outreach
sites becoming the central focus.
The Bradfield maximum cropping method
required complete control of water, both for
irrigation and drainage. Uniformity of slope
across a field was essential. The Landmaster
power source was key to working the soil in
alternative ridges and furrows to achieve ade-
quate tilth, weed control and, most importantly,
water control. Timing of field operations was
During 1972 we learned much from trials
with 'maximum' cropping. First, the ability to
conduct any given operation on time when it
needed to be done was an essential design
consideration. Systems designed for clay soils in
high rainfall areas showed increasing levels of
'management instability' if their soil and water
management requirements were complex8. Our
attempts at maximum cropping in farmers'
fields in two nearby barrios was a near disaster
because of the inability to manage the crops in
a timely way. Component technologies were an
essential building block for productive inte-
grated systems. Locally adapted cultivars and
the technologies to manage them were essen-
tial. A knowledge of specialized techniques for

26 Chapter 2

relay planting and intercropping seemed neces-
sary. Weed and insect control was important.
Economic analysis showed that good crop man-
agement and high yields of a limited number of
crops gave greater net return than wide diver-
sity in inter and relay crops9.
Second, Bradfield's assumptions for the long-
term need for maximum productivity, whether
or not true for the very long term, were not a
useful guide for short-term economic gain for
farmers. Stepped increases in productivity were
more appropriate and economical10. Bradfield
did not anticipate the competing demands for
labour that would actually reduce cropping
intensity in every country but China well before
maximum intensities were achieved. He did
foresee a requirement for mechanization.
Third, the availability and type of power
source was crucial for timeliness of operations,
but costs varied by source. In the early 1970s,
economic conditions in the Philippines were
such that hand and animal-powered operations
were competitive with mechanization for tillage,
despite compromising the timeliness of opera-
tion. It was obvious, however, that power
sources and their cost were critical factors in
systems design11.

2.3.4 National links
It had become increasingly clear that collabora-
tive links with national programmes were
essential for cropping systems development. The
problems of conceptualizing system design, the
complexity of the research process and the
specifics of local environments, all dictated a
need for learning together. There had to be buy-
in from each country into the process.
Collaboration was achieved through training at
IRRI, through many weeks of my own travel
during 1972 and through national multiple
cropping events such as that in Bogor,
Indonesia, in 1973 and in Sri Lanka.
In 1973 the IDRC, the programme donor,
agreed to my argument that we should be
reviewed by our clients, the heads of national
programmes in Asia, on the condition that two
other external reviewers could be named. This
arrangement became pivotal to future pro-
gramme success. Dr David Norman from
Nigeria and Dr Bert Krantz from India joined
what was called the Asian Cropping Systems

Working Group. That group not only reviewed
programme direction, but came to develop
many of the methodologies of the Asian net-
work. The group was to meet twice a year, once
at IRRI and once hosted by a national pro-
gramme. Dr Virgellio (Pexy) Carangal joined the
IRRI team to coordinate the network activities.
In retrospect, the Asian Network, later to
become the Farming Systems Network, was one
of the major contributions of the work of the
early 1970s. At its first meeting at IRRI in
March of 1974 a process for cropping systems
research and extension was hammered out in a
truly democratic fashion.
It cannot be overemphasized how important
this process model was, and the collaborative
way in which it was developed, to the longevity
and impact of Asian farming systems work. The
sequence of observation and description, of
design, test and extension, seems trite today, but
it was critical to methodology developmentl2
and to an orderly sequence of site establishment
and research design. The process model was
supported by recommendations on methods
and organization and on the relationships of
researchers and farmers. Farmer responsibility
for technology access and test became a central

2.3.5 A changing conceptual framework
A conceptual framework for the development of
cropping systems was still not clear. By early
1973 we had discarded the notion of maximum
use of land, water and sunlight. In travelling
across Asia it became clear to me that there
were strong physical determinants of tempera-
ture and water and of soil type to cropping
intensity. We commissioned a group of Asian
scientists to develop an agroclimatic classifica-
tion system for Asian rice-based farming in
1973. Their system was based on monthly
duration of rainfall above and below 50, 100
and 200 mm month 1.13 Two hundred millime-
tres per month for 3 months was deemed ade-
quate for one good rice crop, and for 5 months
for two crops if the first could be direct-seeded.
The Ilo Ilo site in the Philippines was chosen
because of its 5 months of 200 mm month 1
rainfall, and its traditional single crop. Within 2
years the entire area had gone to double-crop-
ping. Maps of climatic zones were subsequently

FSR: Origins and Perspectives

created for Bangladesh, Indonesia and other
countries by L.R. Oldeman. This classification.
and its application to cropping systems work, is
described in detail in Harwood, 1979 (pp.
45-62)14. It eventually evolved into a classifica-
tion of cropping systems research and recom-
mendation domains 5.
A second dimension to systems conceptual-
ization was according to 'development stage', a
dimension classified primarily by a farm family's
degree of participation in a market economy16.
System design changes as cash flow and market
orientation change. Production systems go
through rather predictable changes in structure
as they shift from subsistence-oriented to a mar-
ket economy and then towards production in
support of an industrial economy17.

2.3.6 Modelling controversy
In early 1973 a major controversy engulfed the
programme. A review by the IRRI Board of
Directors made a strong recommendation that
systems modelling be used as a core for the pro-
gramme, and that a systems design expert be
hired. It was recommended that this work be
patterned after the systems modelling work sup-
ported by the Ford Foundation in Thailand and
being used at the International Center for
Tropical Agriculture (CIAT) in Colombia and
the International Institute of Tropical
Agriculture (IITA) in Nigeria. Gordon Banta
and I were adamant in our refusal to take this
direction, for several reasons:
* The three existing programmes were very
expensive, and did not seem to be producing
useful results.
* They used the mainframe computer hard-
ware and software of the day, which had
restricted capability and were unavailable to
non-computer systems scientists.
* Farmers allow no set of parameters to inter-
act over a wide enough range of conditions to
isolate response from 'background' noise.
Farmers, for example, diverted fertilizer to
other crops or switched from planting maize
to other crops at both high and low produc-
tive potential. Models could not be verified
with the computing capabilities of the 1970s.

The controversy, and the pressures for model-
ling, had subsided somewhat by the mid 1970s

as each of the other modelling efforts was
eventually terminated. The commitment
required to oppose allocation of major
resources to premature modelling may have dis-
tracted us too far from developing a conceptual
framework for cropping systems to underpin
modelling research in the future. We focused on
farmer-based testing, perhaps, with hindsight,
to the detriment of the upstream-downstream

2.3.7 Beyond modelling

The need for effective component technologies
quickly became a high priority in the Asian
programme. The availability of high yielding
crop varieties and of adequate seed for them
was of major importance to IRRI's research in
the Philippines and to the network. The region-
wide collection and uniform trials of cultivars
of the major field crops had become a major
activity, and by November 1975 was given
high priority as an IRRI function by the
Working Group I8.
A second area of component technology
work was that of relay and intercropping. A
series of studies carried out at IRRI between
1972 and 1975 quantified the results from
alternative management practices. Relay crop-
ping was the easiest to document, producing
rather straightforward results for the timing of
overlap for various crops and was the most use-
ful information for Asian network collaborators.
Intercrop studies were more complex. We con-
centrated on combinations of maize, mung
bean and groundnut because of their yield
advantage, disease and insect control, and on
maize, upland rice and cassava for its nutrient
use characteristics. Three key points emerged:

* Relay intercrops of 2-3 weeks, under ideal
conditions for stand establishment, do not
effect yields of either crop, but management
of the overseeded crop will often be compro-
* The land equivalent ratio (LER) the
amount of land needed to produce in mono-
cultures the same produce obtained from
one unit of intercrop began to be used in
IRRI literature starting with Bantilan and
Harwood, and Herrera and Harwood, both
in 197320.

28 Chapter 2

* Though an intercrop mix will overproduce
with a land equivalent ratio of up to 1: 6,
the increased production is almost entirely a
result of the longer duration of cropping,
with the period of maximum competition
coming at a time when the lower-canopied
rice crop is least sensitive. Intercrops are
thus not physiologically more efficient than
a well-grown monoculture during the period
when the crops are together21.

Many relay and intercrop studies provided a
wealth of information on the possible role and
weaknesses of traditional practices in modern
systems, as well as an exciting context for grad-
uate research. However, they did not make the
major contribution that cultivar selection, dis-
tribution and testing was to make to Asian agri-
By 1974 the IRRI programme had shifted in
a major way to focus on a few carefully selected
outreach sites in the Philippines, selected
according to agroclimatic zone classification
and farming system type. All had systems ori-
ented towards a market economy and all cov-
ered at least three townships within the same
zone. Research staff took up residence and pro-
tocols were developed for rapid site characteri-
zation, research problem identification and
conduct of trials. As the Asian network was
formed the procedures became more and more
a product of Asia-wide discussion22.
Training programmes began in the late
1960s to train field workers on agronomic tech-
niques for maximum cropping. By 1975 these
had evolved into the IRRI 6-month course for
'site coordinators', covering five general areas;
cropping systems concepts, site selection and
operations, site characterization, economic
analysis and statistics, component production
technologies and extension methods. Degree
training was carried out at the University of the
Philippines, Los Bafios.

2.3.8 Conclusion
Bradfield's contribution to cropping systems
work has been enormous in its demonstration of

the design of integrated systems. The greatest
subsequent impact of the programme of the
early 1970s was its network operation with buy-
in by national programme leaders. Generations
of site coordinators trained in the programme
were to become national research directors in
many countries. Having an IRRI core set of sites,
and intensive crop rotation experiments both at
IRRI and other sites, has provided the backdrop
for training in production methodologies as well
as site and programme management.
With 20 years of hindsight, there was one
major weakness in the programme. We did not
make progress on a conceptual framework for
use in the ex ante determination of systems
change. This would have required more
extended periods of interaction among senior
national research scientists. For such a frame-
work to have been successful, it, like the site
research methodologies which were produced,
would have been collaborative in development.
Improved component technologies, effectively
targeted through systems understanding, will
be the building blocks for future change. The
major resource allocation problem which faced
systems research in the 1970s, as it does now, is
that systems understanding is one step removed
from actual impact, and is not as obvious in its
contribution to eventual improvement. Systems
understanding in turn, requires knowledge of
process-level functions of social, economic, bio-
geochemical and crop and animal physiology
dimensions. Science is moving towards defini-
tions of many of those functions, but our grasp
of many of the key pieces is, even today, quite
Credit for changing rice production in Ilo Ilo.
in the Philippines, from 2 to 10 t ha-1 year-1
went to the newly introduced rice varieties, not
to the systems understanding which brought
them there. This dilemma remains with us 20
years later. But farmers and researchers alike
need such understanding to help guide change
in a world where the pace of change is engulf-
ing us all, sometimes moving us away from the
best interests of farming families, environmen-
tal health and economic stability.

FSR: Origins and Perspectives 29


1. Nye, P.H., 1961. Organic matter and nutrient cycles under moist tropical forest. Plant and Soil, 13,
2. Papadakis, J., 1965. Crop ecological survey in West Africa. FAO, Rome.
3. Ruthenberg, H., 1971. Systems with perennial crops. In: Farming systems in the tropics. Clarendon
Press, Oxford, UK.
4. Bradfield, R., 1964. Some unconventional views about rice culture in southeast Asia. IRRI Seminar, 16
April, 1964.
5. Bradfield, R., 1971a. 1970 Research Program Review, maximum cropping project. 9 Feb., 1971.
6. Bradfield, R., 1970. Increasing food production in the tropics by multiple cropping. In: Research for the
World Food Crisis, American Association for the Advancement of Science, pp. 229-42.
7. Bradfield, R., 1971b. Mechanized maximum cropping systems for the small farms of the rice belt of
tropical Asia. Shin-Norinsha Company, Tokyo, Agricultural Machinery News Weekly and Farming
Mechanization Monthly.
8. Harwood, R., 1979. Small Farm Development: Understanding and Improving Farming Systems in the
Humid Tropics. Westview, Boulder, CO.
9. Banta, G.R., 1973a. Mechanization, labor and time in multiple cropping. Agricultural Mechanization in
Asia, 4(1), 27-30.
10. Banta, G.R., 1972. The outlook for multiple cropping in the Philippines. Presented at the symposium,
'Toward More Progressive Barrios'. April, 1972. University of the Philippines, Los Bailos.
11. Banta, G.R., 1973b. Comparison of power sources in multiple cropping. IRRI Saturday Seminar, August
11, 1973.
12. IRRI, 1975a. Report of the Cropping Systems Working Group. Los Bafios, Philippines. March, 1975.
Harwood, R.R. & E.C. Price, 1976. Multiple cropping in tropical Asia. In: Multiple Cropping. American
Society of Agronomy, pp. 11-40.
13. Coulter, I.K., J.F. Derting, L.R. Oldeman, M.M. Obradovich & T.B. Strattery, 1974. An agro-cli-
matic classification for evaluating cropping systems potentials in Southeast Asian rice growing regions.
International Rice Research Institute. Los Bafos, Philippines.
14. Harwood, R., 1979. (Op. cit.).
15. Garrity, D.P., H.G. Zandstra & R.R. Harwood, 1978. A classification of Philippine upland rice growing
environments for use in cropping systems research. Philippines Journal of Crop Science, 3(1), 25-37.
16. Harwood, R.R. & E.C. Price, 1976. (Op. cit.).
17. Harwood, R.R., 1992. Biological principles and interactions in sustaining long-term agricultural pro-
ductivity. In: Proceedings of the Regional Workshop on Sustainable Agricultural Development. Asian
Development Bank and Winrock International, Manila, Philippines, pp. 2 7-49.
18. IRRI, 1975b. Report of the Cropping Systems Working Group, second meeting. Indonesia.
19. IRRI, 1973. Annual report for 1972. International Rice Research Institute. Los Bafnos, Philippines.
20. Bantilan, R. & R. Harwood, 1973. The influence of intercropping field corn (Zea mays) with mungbean
(Phaseolus aureus) or cowpea (Vigna sinensis) on the control of weeds. Crop Science Society of the
Herrera, W. & R. Harwood, 1973. Effect of intercropping soybean (Glycine max) under different popula-
tions of field corn (Zea mays). Crop Science Society of the Philippines.
IRRI, 1974. Annual report for 1973. International Rice Research Institute. Los Banios, Philippines.
21. Sooksathan, I., 1976. A comparative growth analysis of intercrop and monoculture plantings of rice
and corn. Dissertation. University of the Philippines, Los Bafios.
IRRI, 1977. Annual report for 1976. International Rice Research Institute. Los Bafnos, Philippines.
22. Garrity, D.P., H.G. Zandstra & R.R. Harwood, 1978. (Op. cit.).
Harwood, R., 1979. (Op. cit.).
Garrity, D., R. Harwood, H. Zandstra & E. Price, 1981. Determining superior cropping patterns for
small farms in a dryland rice environment: test of a methodology. Applied Science Publishers,
pp. 269-83.
23. Harwood, R.R. & M. Cavigelli, 1997. Michigan Field Crop Ecology: Managing Biological Processes for
Productivity and Environmental Quality. Extension Bulletin E2646, Michigan State University, East
Lansing. MI.

30 Chapter 2

David Norman

Since the experiment station had been there for almost 50 years, the question that immediately came to mind
was why none of the results or lessons from the research station had 'rubbed off' on neighboring farms.

2.4.1 Introduction

My first professional work experience was as a
full-time researcher linked to the Institute for
Agricultural Research (IAR) at Ahmadu Bello
University (ABU) in northern Nigeria. In 1965,
a rural sociologist, a human geographer and
myself were appointed to staff the Rural
Economy Research Unit (RERU). The Unit had
been set up with a substantial Ford Foundation
grant to catalyse and support interdisciplinary
social science research in the rural areas cov-
ered by IAR's mandate. IAR already had a well-
deserved reputation in technical research and
was well staffed with a highly motivated and
qualified cadre of scientists. The rural sociolo-
gist, B.J. Buntjer, and myself were, however, the
first social scientists appointed by the Institute
and we enjoyed a great deal of freedom in
designing and carrying out a social science
research programme.

2.4.2 Early impressions

In retrospect two important initial impressions
influenced my future career, and hopefully
blunted my potential arrogance as a young,
freshly trained PhD student with all the
answers to the problems of agricultural devel-
opment. The first was surprise at comparing
what was happening on the research station
with what farmers were doing on their own
fields. The experiment station research plots
were largely well tended and weed free; the
crops were invariably vigorous and grown in
sole stands. In contrast, just outside the experi-
ment station fence, farmers' crops were usually
grown in mixtures, were very variable as to
vigour and sometimes suffered from competi-
tion with weeds. Since the experiment station
had been there for almost 50 years, the ques-
tion that immediately came to mind was why
none of the results or lessons from the research
station had 'rubbed off' on neighboring
farms. The second impression was the high
quality of the scientific capacity represented in

IAR, and a strong desire not to alienate these
colleagues. Developing credibility as social sci-
entists within what was historically a technical
research institute, was crucial. We managed to
develop a research programme that was not
confrontational. In retrospect, this was perhaps
lucky. During the 11 years that I was associ-
ated with IAR on a full-time basis there were
never any real antagonisms from the technical
scientists. They were also searching for ways to
make research relevant to the needs of the
farmers in the region. Although I cannot speak
for my colleagues in RERU, I suspect they had
similar impressions, since I do not recall any
disagreements with the approach we laid out
for our work.

2.4.3 A work plan for RERU

The basic work plan which began under RERU
was continued under the Department of
Agricultural Economics and Rural Sociology -
which I headed for about 8 years. It concen-
trated almost exclusively on micro-level studies,
and was divided into four phases as follows':
* Positive phase determining what farmers
are doing.
* Hypothesis testing phase determining why
farmers do things in the way they do.
* Normative phase determining what farm-
ers might do.
* Policy phase determining how the changes
suggested under the normative phase could
be brought about. This also could involve
consideration of the hypotheses tested to
determine whether the suggested policy
changes would conflict with farmers' rea-
sons for doing things in the traditional way.

Much of the research work during the period
1965-71 concentrated on the first two phases.
With the foundation derived from these 'basic
studies' emphasis shifted towards 'change stud-
ies' concentrating on the second two phases.
The experiences in the 'basic study' phases led
to my questioning a 'top-down' approach to

FSR: Origins and Perspectives 31

developing technologies for illiterate limited-
resource farmers. An appreciation of the com-
plexities of developing a process for improved
technologies for such farmers evolved in the
course of the 'change studies'.

2.4.4 Cost-route survey methods
The 'basic studies' were a series of studies in
10 villages in four different areas of northern
Nigeria. Sample farming households were
interviewed twice weekly to obtain a detailed
picture of the stock and the flow of inputs and
outputs relating to the farm family and their
business. Spending considerable amounts of
time in the field with farmers to administer
these very complex and detailed surveys
helped boost the credibility with the technical
scientists in IAR. After some time, however, I
developed three main concerns about the cost-
route methodological approach being used in
the early surveys:
* They were very costly in time and other
resources, not only in terms of the data col-
lection phase but also in terms of analysis.
In fact it took 8 years for the results from one
set of village studies to be finally distributed.
* Much of the data collected were never com-
pletely analysed or exploited. Legitimate
questions could also be asked about combin-
ing in the same analytical framework some
variables measured accurately and others
for which only very inaccurate estimates
were available. Unfortunately, in the field sit-
uation, the criterion for how accurate the
data were was often dependent on how easy
it was to collect, rather than the degree of
accuracy required from the point of view of
* The commitment required for such studies
and to ensure complete farm records, often
meant sacrificing spontaneous interaction
with farmers. Thus the farmer tended to
become an object from which data were
extracted, rather than a colleague from
whom one can learn.

The major limitation of dependence on extract-
ing data from farmers via enumerators was
brought home to me when, after years of
painstaking data collection and analysis, I con-

cluded farmers were rational in growing crops
in mixtures2. I then thought, 'Why don't I ask
farmers why they grow crops in mixtures?'
After 1 week I obtained answers similar to those
from the detailed surveys. Unlike Mike Collinson
who learnt much earlier, it took me some years
to develop a healthy scepticism for detailed farm
management cost-route surveys, and to learn to
place greater reliance on meaningful direct
interaction with farmers. One of my most vivid
memories of this development was a particu-
larly interesting discussion with a farmer who
explained in great detail why he grew certain
crops in mixtures. He had, for example, recog-
nized the beneficial effect of mixing cereals and
legumes, although of course he did not know
about the nitrogen fixing qualities of the latter.
However, his keen visual and mental percep-
tions greatly impressed me, convincing me that
the skills and experiences of farmers should
somehow be put to constructive use in the
search for improvements.
Important lessons were learned during the
data collection and analysis stages which have
had a major influence on my evolution towards
advocating a farming systems approach. Four of
these lessons were:
* It is vital to try to understand the farming
system with which one is working. This can
only be done in the field through interacting
with those responsible for operating that sys-
tem. If this is not done then there is a real
danger of misleading interpretations with
'outside' eyes.
* There was great heterogeneity in the com-
plex farming systems being implemented
because of differences, not only in the bio-
physical factors, but also in the socioeco-
nomic factors, faced by different farming
households. The strategy of recommending
blanket technology packages did not, there-
fore, seem relevant.
* I also developed a healthy respect for the
farmers' skills in operating in such complex
production environments and for the ratio-
nality of their behaviour. Examination of
issues relating to mixed cropping in particu-
lar, influenced my rapidly developing respect.
* Because of the seasonal nature of rainfed
agriculture in northern Nigeria, labour
shortages at certain times of year were the

Chapter 2

norm. At the same time, few improved tech-
nologies were used. Access to, and availabil-
ity of, purchased inputs was very limited.
Given these constraints and realities were
the technologies being recommended at the
time likely to be relevant?

2.4.5 Evaluating technologies with farmers
The knowledge and impressions built up during
the 'basic studies' led naturally to the next step
of evaluating, with farmers, some technological
packages being recommended by extension
from work done at IAR. This involved farmer
volunteers testing technological packages for
sorghum, maize, cowpeas and cotton, in a
farmer managed and farmer implemented
(FM/FI) mode. Detailed records were kept as
farmers grew test plots and compared their
results with their own adjacent control plots.
Again this was a major learning experience
and, in addition to reinforcing many earlier
insights, produced new ones.
Very few farmers tested the packages in the
way researchers had intended and many devi-
ated significantly from the recommended meth-
ods. Sometimes they were unable to do exactly
what was recommended because, for example,
labour was not available for weeding when it was
required. However, deliberate decisions were
made to deviate, with some giving priority to get-
ting food crops established before planting cot-
ton, for example. Because of these deviations I
was constantly consulting with technical scien-
tists on the next best alternative for the farmers.
For example, should a top dressing of fertilizer be
put on the field if the weeding operation was
inadequate? This later led me to argue for recom-
mendations to specify the best stepwise adoption
if a technology package is involved. I also argued
that recommendations should include condi-
tional and targeting information giving guide-
lines about what would be best to do if deviations
occur, and for what types of situations the rec-
ommended technology would be most suitable.
The objective was to make the recommended
technologies at least partially relevant to as
many farmers as possible, and to exploit to the
maximum extent, the expertise and knowledge
of the technical scientists themselves.
The evaluation criterion in developing the
technologies on the experiment station centred

on yields per hectare which was often incom-
patible, or at least not completely congruent,
with the criteria used by farmers in their fields.
Given the importance of labour as a resource in
the seasonal agricultural cycle, the package
only increased the value of returns per unit
area and per unit of labour used when applied to
maize not a major traditional crop. The other
packages increased the net return per unit area
but not the return to labour3. In retrospect, it is
not surprising that later, when the World Bank
appeared with their integrated rural develop-
ment projects with efficient input distribution
and product marketing systems maize areas
expanded and production increased dramati-
cally4. These area increases occurred largely at
the expense of sorghum and even legumes. The
obvious lesson from this was that irrelevant
evaluation criteria on experiment stations
inevitably lead to the recommendation of irrele-
vant technologies.
Given the somewhat negative connotation of
such findings there was the possibility of con-
frontation between the social and technical sci-
entists. Given the years of effort the technical
scientists had spent developing the technologies
this would be understandable. To safeguard
against possible confrontation, efforts were
made to consult technical scientists before the
testing programme was implemented. The most
rewarding experience was working with the
cotton scientists who showed a definite desire to
collaborate actively to test their technological
package with farmers. This, as it happened, was
very significant because the recommendation
called for early planting of cotton to maximize
cotton yields. Within 1 month of collaborating
in the field the cotton scientists realized this was
an unreasonable expectation. This led to very
constructive collaboration between the techni-
cal and social scientists. Consequently, the
potential value of later planted cotton was
examined, together with replacing the water
demanding and labour intensive spraying sys-
tem with an ultra-low volume (ULV) (i.e. oil-
based) insecticide spraying regime5. These
experiences taught us to consult before embark-
ing on something that might be sensitive to oth-
ers and, if possible, to collaborate actively with
the other parties so that the eventual findings
are more acceptable. If progress in meeting the
needs of farmers was to occur, I became

FSR: Origins and Perspectives

increasingly convinced that technical and social
scientists had to collaborate. This also stemmed
from the realization that evaluating anything
effectively on the experiment station that was
not technical in nature was difficult for scien-
tists. Evaluating yields in terms of per unit of
area is, therefore, much easier than in terms of
per unit of labour.
Perception of the huge gap between what
technical scientists do and can do on the experi-
ment station and what farmers do, and can do,
on their farms, led to a conviction that major
changes were necessary in the way in which
technologies were designed and evaluated. This
conviction became focused on the notion that
continued ex post evaluation of technologies by
social scientists was not very efficient, and
could, in fact, be considered counterproductive.
Instead, attention needed to be focused on inter-
disciplinary collaboration between technical
and social scientists to develop and evaluate
technologies in more of an ex ante operational
mode. It was also becoming obvious to me that
farmers had to become actively involved in that
process, because of the intimate knowledge
they possessed about local production environ-
ments and the fact that they could actively and
constructively contribute to the development
and evaluation of relevant technologies. These
convictions were supported and nurtured by
two other influences. One was a growing
awareness that workers in other areas were also
going through something of the same transi-
tion. For example, I remember being particu-
larly impressed with some of the work of the
Unite Experimentales in Senegal with its philos-
ophy that one learns through perturbingg' the
system. And in the early 1970s I became a
member of the Asian Cropping Systems
Network, coordinated by IRRI, which resulted
in very useful trips to Asia where the same
issues were under discussion.

2.4.6 Reflections
Looking back, the 11-year experience at IAR
was truly rewarding and I now realize that, in
many ways, what was happening there was

ahead of its time. There was little confrontation
between the social and technical scientists and
the working climate was open to changes.
There were, for example, formal annual meet-
ings between scientists and extension staff to
discuss research programmes and results.
During the early 1970s the research pro-
grammes were reorganized into multidiscipli-
nary commodity/subject matter teams
(including a farming systems-based team), and
experiment station research work started on
mixed cropping. The working environment into
which the social scientists were introduced in
the mid 1960s was, therefore, supportive rather
than antagonistic, enabling their contribution
to be nurtured a unique opportunity for those
of us there at the time.
Perhaps the final step in embracing the
farming systems approach came from a Ford
Foundation farming systems workshop hosted
by the Institut d'Economie Rurale in Bamako,
Mali, in mid 1976. That workshop was signifi-
cant for at least three reasons. The first was that
it accepted that the description of farming sys-
tems should be viewed as a means to an end
and not an end in itself. The second was that the
workshop developed and approved an analytical
framework6 to improve the efficiency with
which relevant improved technologies are
developed and evaluated. That analytical frame-
work, still in use today, was first drawn on a
bedsheet, compliment of the Hotel de l'Amitie!
The third was that this was called 'the farming
systems research approach', the first time I per-
sonally recall this label being used.
Obviously since then, as a result of inter-
acting with many farming system practition-
ers, nearly 9 years of further field experience
in Botswana, and many short-term assign-
ments in different parts of the world, my per-
sonal evolution in the application of the
farming systems approach has continued.
However, the basic principles that I embraced
during the early years of my career are still as
valid today. Perhaps the most important of
these principles has been the belief that the
farmer should be, and has a right to be, at the
centre of the stage.

34 Chapter 2

1. Department of Agricultural Economic and Rural Sociology, 1976. Progress Report no. 10. Zaria,
Institute for Agricultural Research, Ahmadu Bello University.
2. Norman, D.W., 1974. Rationalising mixed cropping under indigenous conditions: the example of north-
ern Nigeria. Journal of Development Studies, 11(1), 3-21.
3. Norman, D.W., E. Simons & H.M. Hays, 1982. Farming Systems in the Nigerian Savanna: Research
and Strategies for Development. Westview Press, Boulder, CO.
4. Smith, J., A.D. Barau, A. Goldman & J.H. Mareck, 1994. The role of technology in agricultural inten-
sification: the evolution of maize production in the northern Guinea Savanna of Nigeria. Economic
Development and Cultural Change, 42(3), 537-54.
5. Norman, D.W., J.A. Hayward & H.R. Hallam, 1974. An assessment of the cotton growing recommen-
dations as applied by Nigerian farmers. Cotton Growing Review, 51, 266-80.
See also:
Norman, D.W., P. Beeden & J.A. Hayward, 1976. A comparative evaluation of ultra-low volume insec-
ticide application on cotton farms in the north Central State of Nigeria. Nigerian Journal of Plant
Protection, 2, 23-9.
6. Gilbert, E.H., D.W. Norman & F.E. Winch, 1980. Farming Systems Research: a Critical Appraisal. MSU
Rural Development Paper no. 6. Michigan State University, East Lansing.

Mike Collinson

My involvement with the scientists on the research station impressed on me the overwhelming need to
understand small farmers whose objectives clearly differed from those of crop researchers. They also differed
from the objectives of commercial farmers, whose motivations underpinned the conventional western
approaches to farm management.

2.5.1 Introduction

Two long periods of my professional life were
devoted to FSR. The first, from 1960 to 1971,
including nearly 10 years in Tanzania, focused
on understanding the adoption of technologies
by smallholders. The second, from 1975 to
1987, based in Kenya with CIMMYT, focused
on institutionalizing FSR-based on-farm
research in eastern and southern Africa. The
first phase provided the conviction for the sec-
ond, so I dwell on the 1960s in this description.
In 1961 I started my first career job at the
Western Region Research Centre, Ukiriguru, in
the north-west of Tanzania then Tanganyika
- joining a dozen 'hard' scientists; in plant
breeding, agronomy, pathology, entomology
and soil science, all trying to raise crop yields.
Most were working on cotton, the main
regional cash crop, others on important food
crops, such as maize and rice, which were also
sold in local markets. My task as the first social
scientist on the research station (and, as it
turned out, the first in the agricultural research
institutions of East Africa) was to document the

priorities of the farmers of the region and how
they used their resources to meet them.

2.5.2 Professional baggage

I brought to the work a professional baggage
acquired in UK universities. This baggage
included the research and extension approach
then used in the UK Farm Management Service
in which constraints on the performance of the
individual farm were identified by comparing
data for that farm to standards for a group of
farms of the same type1. In the late 1950s
whole farm planning had just found a niche in
the farm management arsenal. In 1958 in the
USA Heady and Candler' detailed the applica-
tion of linear programming (LP) to farm plan-
ning. Five years later, Clayton provided an early
example of its use in peasant agriculture3.
Both approaches involved a farm manage-
ment professional in one on one contact with the
individual farmer, to collect data for analysis
and to give advice. Earlier work in farm manage-
ment in developing countries was limited. In
1953 Jolly began to plan and develop holdings

FSR: Origins and Perspectives

for small resource-poor farmers by helping
selected families to build up 'Unit Farms' under
controlled conditions at the college campus in
Trinidad. Surveys of small farmers in the 1950s
were being pioneered by Morgan-Rees in
Zambia and Edwards in Jamaica; the only for-
mal attempts I was aware of to bring farm man-
agement principles to the improvement of the
small-farm sector in developing countries.
Beyond conventional farm management, stud-
ies by anthropologists such as Richards in
19394, Conklin in 195 75, and De Schlippe and
Batwell in 19556 had convinced me of the
value of understanding small farmers' systems.
Jolly's work in Trinidad also proved particularly
valuable to my research in Tanzania in an
unexpected way and I return to this.

2.5.3 Conventional methods in crop
My involvement with the scientists on the
research station impressed on me the over-
whelming need to understand small farmers
whose objectives clearly differed from those of
crop researchers. They also differed from the
objectives of commercial farmers, whose moti-
vations underpinned the conventional western
approaches to farm management. Three aspects
stood out:
* In the early 1960s, and in many communi-
ties today, cash income could not guarantee
command over food supply. For most families
food marketing was essentially local. In a bad
year there was scarce food at home and noth-
ing to buy in the local market. Family priori-
ties were for a combination of foods produced
on the farm itself that was reliable day in day
out. These food security priorities dominated
farm decisions on resource allocations.
* Small farmers managed the full range of
external uncertainties; weather variation,
degrading land, pests and diseases and unre-
liable markets and prices, not through the
protective devices available to commercial
farmers, such as insurance, but through
their own enterprises and husbandry prac-
tices. Fully exposed to this diversity of uncer-
tainties they used the production strategies
evolved by their culture for survival.
* Small farmers using hand hoes, and even
those using oxen for ploughing, have very

limited power at their disposal. The returns
to spending limited cash on increased power
to extend the area cultivated were, and are,
often better than the returns on purchasing
inputs to intensify production on the exist-
ing area, especially given the vagaries of
input supply, and the physical difficulties in
These three aspects in particular caused me to
reflect on the commodity approach to crop
improvement on the research station. The sci-
entists there, in the classic reductionist tradi-
tion, were trying to identify the best germplasm
and best husbandry to maximize yields of their
commodity in the conditions particular to west-
ern Tanzania. Many characteristics of classical
experimental methods and of research station
operation isolated their results from the real
world of the small farmer, and the current prac-
tices of those farmers were never the context in
which the recommendations were identified.
The criteria used by farmers in evaluating rec-
ommendations were rarely compatible with
higher physical yield per unit area of land
monopolizing the attention of the scientists.
Heavy machinery on the station meticulously
prepared the fields, creating a tilth impossible
for the hoe to achieve. Machinery meant that
acres could be prepared and sown after each
planting rain, while farmers might be limited to
the half acre their family could prepare and sow
in the 2 or 3 days following that rain. Unlimited
and unrecorded labour tended the experimental
plots and high levels of non-treatment variables
allowed the full expression of differences
between treatments, yet created a multi-compo-
nent technological package. Both non-treat-
ment and treatment variables were usually new
technology components as far as the farmers
were concerned. Each one used their labour or
cash in new ways and, through clashes with
other enterprises competing for these resources,
often jeopardized the family's priority an
assured, preferred food supply.
There was a stark contrast between the
sophistication of design and statistical analysis
for experiments, the ad lib resources used for
their management, and the all too common
irrelevance of the criterion for evaluating the
results. Cotton research dominated the station
and had first priority for equipment as the sea-
son started. Because of the threat of American

Chapter 2

bollworm building up on early-planted maize
and migrating to the cotton crop, maize plant-
ing experiments on the station were delayed
until some 30 days after cotton establishment.
Farmers, in contrast, wanted a supply of new
season food flowing from the farm as soon as
possible. They enjoyed fresh maize, and, in years
following bad harvests, their only option was to
buy it locally at four times the normal price. I
tried to highlight this kind of dilemma in an
article I wrote in 19687 focusing on the then
current recommendation to farmers to use their
limited cash to purchase fertilizer for cotton. I
argued that farmers in western Tanzania would
gain more from fertilizing their maize and
growing a smaller area to meet their fixed
needs, thereby releasing labour to plant a larger
area of earlier cotton. The systems perspective
was nicely illustrated; the benefits from putting
fertilizer on maize came from a larger acreage of
cotton, and an increase in yields from planting
some of it earlier.
The whole issue of evaluation criteria
among small semi-subsistence farmers demon-
strates a gross irrationality in the application of
science which unfortunately still prevails. The
social sciences remain a rarity in most agricul-
tural research institutions. Where they have
found a niche they are still fighting an uphill
battle to influence the technical establishment
to modify the traditional research process.

2.5.4 A trial farm
In 1962, following Jolly, I established a 'trial
farm' on the research station operated by a local
family and managed between the farmer and
myself with advice from the station scientists8.
It proved to be the greatest single influence in
changing station scientists' attitudes towards
small farmers and their systems. In the first
year I discussed with the family their food pref-
erences and the quantities they would need to
feed them through the year. They would grow
maize, rice, cassava, sweet potato, cowpea,
groundnut and pumpkins for food, and cotton
for cash. I turned to the scientists for the best
way to grow these crops. The most striking
aspect of their recommendations was that every
commodity researcher expected his crop to be
planted immediately after the onset of reliable
rains typically the first week in December. I

pointed out that it required some 18 man days
to prepare and plant an acre of seedbed with
hoes. With less than two labour units, to pre-
pare and plant the 3 acres needed for food crops
would take 3-4 weeks and the 2.5 acres of cot-
ton planned would also take 3 weeks. I asked
the scientists to agree which crops the farmer
could delay planting with least compromise on
yield potential. Their responses were all the
same: 'any other- but not mine!'
The trial farm adopted farmers' priorities as
a starting point. During its 5 year life it
required huge compromises on best practice for
any single commodity in order to accommodate
the farmers' cash and labour resources.
Nevertheless, in its first 3 years it provided
returns to available labour 250-300% higher
than those of the average local farmer and
150-200% higher than those of the best 30%
of local farmers. It was an education for me and
for all the scientists on the Centre. The com-
promises, however, caused reflection and exten-
sive controversy on future priorities:
* Which crops had the greatest tolerance for
delayed planting?
* Was there diversity in varietal tolerance to
delayed planting within species which the
breeders could exploit?
* Identifying shorter-term materials for crops to
be planted later in the 6-month rainy season.
The strong and well-established interactions
between time of planting, variety, plant density
and fertilizer levels also demanded a re-exami-
nation of husbandry practices for crops for
which planting was delayed.
It was years before these research themes
were taken up, for example, with maize, sun-
flower and cassava in eastern and southern
Africa. Even today it remains difficult for many
plant breeders to accept that it may be better for
the farmer to grow a maize with a shorter
maturity period and lower yield potential:
* If it allows new food supply earlier in the
season from the farm.
* If it allows a second crop to be planted.
* If it allows an extra field to be planted to
maize months after the start of the rains
when earlier planted fields have suffered for
one reason or another.

FSR: Origins and Perspectives

2.5.5 Scarce professionals and countless
I started by applying the cost-route approaches
to survey learned in university but was brought
up short by early experiences in my own work
and exposure to wider efforts to develop small-
holder agriculture in the early 1960s. An early
lesson came from the breakdown of the Village
Settlement Scheme introduced by the
Government of Tanzania and the World Bank in
1964. An International Bank for
Reconstruction and Development (IBRD) report
to the Tanganyika Government in 1960 listed
the entire graduate strength of the Tanzanian
extension services as 58 people, and advocated
new approaches to improve their effectiveness.
The scheme which emerged proposed 69 settle-
ments with an average of 250 families in a vil-
lage and machinery provided to work the
2500-3750 acres planned for growing village
crops. After implementing seven of the settle-
ments it was clear that the management needs
of 69 such villages would drain all the graduate
and diplomat staff from the ministries' exten-
sion services. This scheme would have focused
the nation's entire professional manpower on
some 15,000 families a small proportion of
even the annual increase in the rural popula-
tion. Further, the special circumstances within
these schemes, particularly the full tractor
mechanization of cultivation, implied that the
innovations would be irrelevant to the rural
population at large.
My involvement with these schemes crystal-
lized a key dilemma in evolving an approach to
small-farmer improvement. The scarcity of pro-
fessional manpower and the large numbers of
small farms precluded one on one interaction
between professional and farmer. It ruled out
the conventional approaches used by farm
management professionals for farm family
income improvement in the USA and the UK.
The large number of farmers, the diversity of
farms and their small size with low levels of pro-
duction, limited the returns from investment in
professional time to improve individual farm
units. This pushed me towards faster and
cheaper methods of data collection. I was seek-
ing wider coverage with sufficient accuracy to
understand farmers' aims and the way they
used their resources to achieve these. The
search had four phases.

2.5.6 Lowering the cost of understanding
small farmers

An initial concern was with stratification.
Common practice in survey work in the early
1960s was to stratify on the basis of adminis-
trative areas; districts or ginnery zones; units
for which official decision making was already
institutionalized. In western Tanzania most
farming systems were providing full family
subsistence. However, the dominant starch
staple differed widely: maize, plantain, cas-
sava, rice or sorghum, from one area to
another. The main starch staple usually occu-
pied as much as 70% of the area cultivated
and absorbed most of the labour. The crop cal-
endar, cultural practices, and therefore
resource demands, differ across these staples
and each system had to be sampled separately
to avoid confounding an understanding of
farmers' priorities, strategies and production
decisions. This began a search for cost-efficient
ways to identify a typology of farms.
An early preoccupation in data collection
was to get away from expensive cost-route stud-
ies in which the measurement of production,
expenditure and labour use dominated survey
design and analysis, requiring frequent visits to
each farm in the sample to collect daily data.
The issue was the trade-off between visit fre-
quency and data accuracy. Sixty observations
within a farming system consistently gave an
acceptable sampling error which could be man-
aged with a degree of information about the
population being surveyed the real issue was
observational or enumerator error.
In the course of some 20 field surveys on
small farmers between 1961 and 1966, I1 incor-
porated a series of data collection experiments
for important parameters. Collection intervals
ranged from daily to yearly visits by enumera-
tors and different intervals were used with dif-
ferent samples, sometimes subsamples from the
same population. My guidelines for the experi-
ments came from Zarkovich9 whose later char-
acterization of data as 'events'"1 brought clarity
to a complex issue. He drew a distinction
between regular and irregular events, and
between frequent and infrequent events. He
concluded, particularly for regular events, that
respondents often answered out of experience
rather than recalling the specific event.
Exploiting this idea, I paid more attention to

38 Chapter 2

labour data. All new technologies, through
changes in timing or the intensity of labour
required, have important implications for
labour redistribution. Consequently they impact
the priorities of farm families which dictate the
existing patterns of labour use. Data collection
experiments showed that drawing on farmers'
knowledge gave rates of work per unit area on
the main crop operations not significantly dif-
ferent from labour rates based on frequent visit
data collection techniques11.
In the single visit method each main enter-
prise was covered separately. The farm family,
with the women responding for the operations
for which they were responsible, took the sur-
vey enumerator to a field previously planted
with a crop for which labour data was needed.
The enumerator measured the field. Standing in
the chosen field, families were first asked to cal-
endar the operations done on the specified crop
in this field in a typical year. Then a sequence of
three questions was asked about each operation
from the beginning to the end of the season:
* Which members of the family, and/or hired
labour, would normally work on this opera-
* For each worker, what time would work
start, what time would it finish, on a typical
day while working on this operation?
* With this group of people working, how
many days would it take to finish the opera-
tion on this field?
The method provided labour data in a single
visit to the farm. Variation was relatively low,
and 30 observations within a farming system
gave sufficient accuracy, allowing coverage of
the five or six main enterprises by enumerating
two or three on each of the 60 units within the
sample. The data were used to build a labour
calendar for the farming system to evaluate the
impact of new technologies on labour distribu-
tion and on family priorities, including food
supply, reflected by this allocation of labour.
Later work for the World Bank12 confirmed that
farmer estimates, properly enumerated, are as
accurate as crop cutting for the measurement of
production levels.
Despite relatively cheap collection tech-
niques, returns to labour are still largely
ignored in evaluating potential innovations.
The farmer's bottom line is a reliable, year in

year out, return to household labour, even
where land is very limited.
As survey experience accumulated it became
clear that, within a farming system, there were
many attributes common to the population as a
whole. Often culturally determined, these
attributes could be investigated by discussion
with individuals, or with groups of farmers.
Describing these attributes gave an understand-
ing of the important dimensions of the system,
as well as insights for better planning of more
formal surveys. The 'pre-survey' evolved'3.
Informal but carefully organized interactions
with relatively few farmers provided qualitative
information on a wide range of attributes of the
farming system. These included the crop calen-
dar, husbandry methods, rotational practices.
changes in enterprises, land acquisition and
tenure rights, preferred dishes and their substi-
tutes, seasonal eating patterns and obligations
to the community with respect to land, labour
and livestock. This pre-survey evolved into the
informal survey which was developed in FSR to
provide a qualitative understanding of the farm
as a whole, often as the foundation for soft sys-
tem modelling14.

2.5.7 Technology adoption by small farmers
In 1970. after almost 10 years' work in
Tanzania, I went back to the UK. to the
University of Reading to do a PhD. Using a lin-
ear programming package I modelled a farming
system for which I had 3 years of detailed data.
The 32K Elliot computer used batch processing
and occupied two rooms in the applied statistics
department very different from today's laptop!
My initial aim was to let traditional farm enter-
prises and improved enterprises, based on
research results, compete for farmers' resources
to best satisfy family food and cash priorities:
* Preferred constituent foods to be mixed in
dishes favoured in different seasons.
* Preferred, in-season fresh foods supplied
direct from the field.
* Insurance foods (cassava, sweet potato)
grown to manage the risks of preferred food
* Cash.
Optimizing the model offered a 300% rise in
cash income. Yet it was clear that the manager-

FSR: Origins and Perspectives 39

ial changes in the farm system required for this
improvement were hugely complex and, given
the limited information on season-to-season
replicability of the results of the improved tech-
niques, the results were also uncertain. For me
the revelation was the fact that although I now
had a target, an optimal system which looked
good, I also had two far more complex questions
to work on:

* What sequence of innovations would allow
farmers to improve on their current system,
move towards the target system, and mini-
mize the risks of family food scarcity?
* Would the target remain the same as farm-
ers progressed towards it?

Acknowledging the dynamics of both food and
cash-crop markets, the answer to the second
question had to be no.

2.5.8 Conclusions

That work left me with the conclusion that the
optimal system is an illusion. The key chal-
lenges were to identify the immediate steps for-

ward that would be most acceptable to farm-
ers15, identify homogeneous groups operating
the same system to use scarce R & D resources
efficiently, and to find cost-effective methods in
understanding farmers' systems.
I came to FSR as a result of the shortcom-
ings of my original professional baggage when
faced by the circumstances of African small-
holders. By the early 1970s I was confident the
approach would have an important future16
and determined, under a CIMMYT umbrella, to
promote an FSR-based approach to technology
development in eastern and southern Africa.
Recalling this history brings me back to a
question posed in introducing the book: over a
generation later how do African, Asian and
Latin agricultural professionals differ in the
tools they now bring to the job? Furthermore,
are they better equipped? If so, can some of the
improvement be attributed to the FSR move-
ment? Where today's professionals are not bet-
ter equipped, does the problem lie in FSR itself
or in prevailing institutional conditions? Many
of the other contributions in this book shed
light on this important and ongoing issue.

1. Blagburn, C.H., 1961. Farm Planning and Management. Longmans, London.
2. Heady, E.O. & W. Candler, 1958. Linear Programming Methods. Iowa State University Press.
3. Clayton, E., 1963. Economic Planning in Peasant Agriculture. Department of Economics Monographs,
University of London, Wye College, UK.
4. Richards, A.T., 1939. Land Labour and Diet in Northern Rhodesia. OUP Oxford.
5. Conklin, H.C., 1957. Hanunoo Agriculture: a Report on an Integral System of Shifting Cultivation in the
Phillipines. FAO, Rome.
6. De Schlippe, P. & B.L. Batwell, 1955. Preliminary study of the Nyangwara system of agriculture in the
southern Sudan. Africa, XXV(4), 321-350.
7. Collinson, M.P., 1968. The evaluation of innovations in peasant agriculture. East African Journal of
Rural Development, 1(2), 50-9.
8. Collinson, M.P., 1969. Experience with a trial management farm in Tanzania, 1962-65. East African
Journal of Rural Development, 2(1), 28-43.
9. Zarkovich, S.S., 1961. Sampling for Censuses and Surveys. FAO, Rome.
10. Zarkovich, S.S., 1964. The Quality of Sample Statistics. FAO, Rome.
11. Collinson, M.P., 1972. Farm Management in Peasant Agriculture. Praegar, New York.
12. Murphy, J., D.J. Casley & J.J. Curry, 1990. Farmers' estimations as a source of production data. World
Bank Technical Paper 132. World Bank, Washington, DC.
13. Collinson, M.P., 1972. (Op. cit.).
14. Rhoades, R., 1981. The Art of the Informal Agricultural Survey. Training Doc. 1982-2. Social Science
Department, CIP, Lima.
15. Collinson, M.P., 1972. (Op. cit.).
16. Collinson, M.P., 1974. Transfering technology to developing economies: the example of applying farm
management economics in traditional African agriculture. World Development. 2(2). 93-7.

Chapter 3

FSR Understanding Farming Systems

Robert Hart

If concepts are generalizations, a conceptual framework is a set of interconnected generalizations.
Agreeing on a conceptual framework is one of the most difficult aspects of interdisciplinary research,
and farming systems research is no exception.

3.1.1 Introduction
Biological scientists, social scientists, farmers
and policy makers all view the same reality
from different perspectives. When farming sys-
tems research (FSR) was just getting started,
newly formed interdisciplinary teams spent a
long time agreeing on a common conceptual
framework. Researchers who followed used the
existing frameworks as a starting point and
adapted them. Each new initiative brings
together a new team that is never completely
satisfied with past efforts, and a new conceptual
framework is developed.
This chapter is a subjective retrospective,
analysing the evolution of the conceptual
frameworks that guided the research and
development processes I observed at first hand.
It is biased by my own disciplinary back-
ground in ecology and agronomy and, of
course, by the approaches taken by the inter-
disciplinary teams with which I have worked.
Specifically I should note the influence of
colleagues at the Centro Agronomico Tropical
de Investigation y Ensensana (CATIE), the
Caribbean Agricultural Research and
Development Institute (CARDI), Winrock
International and the Rodale Research Center.
The conceptual frameworks that my
colleagues and I constructed were strongly

CAB International 2000. A History of Farming
Systems Research (ed. M. Collinson)

biased towards an ecological perspective. The
frameworks we constructed were never static
and were constantly evolving, even over the
life of a given project.
Three general tendencies are obvious:

* There has been an expansion in the scale of
the target systems from an early emphasis
on crop populations towards a later interest
in farm system and watersheds.
* There has been an expansion in the criteria
used to evaluate system performance from
an early emphasis on productivity towards a
later interest in stability and criteria related
to sustainability.
* There has been an expansion in the target
beneficiaries from an early emphasis on
'small farmers' towards a later emphasis on
women and gender issues, urban as well as
rural poor and the current interest in sus-
tainability and the benefits that will be
received by future generations.

These three tendencies are summarized in
Table 3.1.1 and described in detail in the first
three sections of this chapter. In the fourth
section I have tried to analyse the operational
implications of this expansion of the concep-
tual framework used to guide the different FSR


42 Chapter 3

Table 3.1.1. Changes in the FSR conceptual framework.
1970s 1980s 1990s
System scale Cropping systems Cropping systems Cropping systems
Livestock systems Livestock systems
Farming systems Farming systems
Community systems
Performance criteria Productivity Productivity Productivity
Stability Stability
Target beneficiaries Small farmers Small farmers Small farmers
Women Women
Next generation

3.1.2 Expanding target systems

FSR has many mothers. Social scientists, ecolo-
gists and agronomists all have a legitimate
claim to having played a role in its conception.
Some people are likely to cite the influence of
anthropologists and agricultural economists
who in the 1960s highlighted the fact that
many traditional farmers manage complex
multi-species cropping systems for example
Norman, 19681. Others will note the role of
ecologists and agroecologists who pointed out
the virtues of higher diversity Margalef, for
example, in 19682.
Agronomists working with perennial crops
and pasture species had been interested in multi-
species cropping systems for many years. In the
1960s and early 1970s there was a dramatic
increase in interest in annual crop-based multi-
species cropping systems and hundreds of agro-
nomic papers on intercropping were published.
But financial support for this type of research
was, in general, limited because the multi-
species cropping systems used by resource-poor
farmers were viewed by many mainstream
development institutions as part of the problem,
contributing to low farm-level productivity. This
changed dramatically when Richard Bradfield
began his research with multiple cropping sys-
tems at the International Rice Research Institute
(IRRI)3. His prior international reputation as a
soil scientist helped to legitimize research on
multi-species cropping systems and donors
began to provide the type of financial support
needed for interdisciplinary research.
In the 1970s most cropping systems teams
were looking at two-crop systems like

rice/wheat rotations and intercropped maize
and beans. Today the emphasis is on the analy-
sis of complex watershed-level systems. This
expansion in FSR's conceptual framework was
no orderly chronological expansion in the limits
of the systems under consideration. Naturally,
while many people were doing plot-level
research with 'simple' cropping systems, others
were analysing watershed hydrology and
regional land use patterns. But in most cases.
even though this research with larger systems
was going on concurrently, FSR teams doing
the plot level on-farm research in the 19 70s did
not include these larger systems within their
conceptual frameworks.
But the expansion in system scale was not
completely unsystematic. A common character-
istic of the conceptual frameworks of many
cropping system research initiatives was an
emphasis on the hierarchical relationship
among systems. This seems to have occurred
independently in different countries unsur-
prising given the importance of the concept of
hierarchy in systems theory. The authors who
most stressed the importance of hierarchy as an
organizing principle were, perhaps, Gordon
Conway4, Louise Fresco5 and myself6. Many
cropping systems teams conceptualized crop-
ping systems as subsystems of farms, that
could, in turn, be viewed as subsystems of com-
munities or watersheds or subregional systems
functioning within the larger national and
global systems (see Fig. 3.1.1). It was almost
inevitable that once researchers started to see
the importance of understanding the suprasys-
tem in which their target system functioned,

FSR Understanding Farming Systems

farming systems

cropping systems

Fig. 3.1.1. FSR's expanding target systems: FSR's conceptual framework has expanded from an initial
emphasis on cropping systems in the 1970s, to an emphasis on farming systems in the 1980s, to an empha-
sis on watershed-level systems in the 1990s.

they would begin to 'climb' the systems hierar-
chy ladder in their quest to understand the
environment determining the structure and
function of their target systems.

Cropping systems
Even after many years of cropping systems
research, it is still not always clear how different
FSR teams defined a 'cropping system'. The two
most common approaches were:
* To limit the components of a cropping sys-
tem to the crop plant populations that inter-
act in space or time.
* To include within a cropping system, in
addition to the crop populations, the soil and
soil organisms, weeds, insects, pathogens,
etc., that interact with the crops.
The first approach evolved mainly from crop
physiology and autecology. The second evolved
out of ecosystem ecology.

Prior to the evolution of cropping systems
research, the conceptual framework for agricul-
tural research and development was strongly
influenced by plant breeders. Agronomists
manipulated the environment so that new vari-
eties could take advantage of their 'improved'
genetic potential. Situations that undermined
this, such as competition from other crop popula-
tions, low soil fertility and pressure from herbi-
vores, were to be avoided. Cropping systems
research turned these concepts upside-down. Its
proponents suggested that cropping systems
should be improved by changing the arrange-
ments of crop populations in space and time and
by using varieties that could fit into these pat-
terns. They also suggested that genetic improve-
ment should be directed towards developing
germplasm that could be used to develop more
productive cropping systems rather than towards
the development of higher yielding varieties.

Chapter 3

Farming systems
Once efforts were underway to develop better
cropping systems and better agricultural
ecosystems, teams soon began to expand
research to include interactions between crops
and livestock and between crops and trees. This
tendency was strongly influenced by the
increasing emphasis on on-farm research sup-
ported by complementary field station experi-
ments. It became clear that farmers did not
manage cropping systems in isolation; they
manage farms in which the cropping systems of
interest to the researchers are only one of many
farm subsystems.
Animal science has a long tradition of taking
a systems approach to the improvement of live-
stock production. While agronomists were mov-
ing towards on-farm research and farm-level
analysis, animal scientists in many countries
were developing production modules integrating
feed production with herd management.
However, the merger of agronomic and animal
science approaches was hindered by the fact that
agronomists were working with smaller farmers
and animal scientists tended to work with
medium- to large-sized farmers and ranchers. In
the 1980s, agronomists and animal scientists
had to agree on a common conceptual frame-
work in order to work on issues such as the use
of crop residue as livestock feed, and animal
traction and manure for cropping systems.
The evolution from crop/livestock systems to
an interest in improved farming systems
changed the relationships between agronomists
and animal scientists with economists.
Economists had always played a key role on
both cropping systems and animal production
systems teams, but this was primarily to use
data gathered by biological scientists for eco-
nomic analyses. The economists rarely had the
resources to do much more than make brave
assumptions about labour use, opportunity
costs of inputs, and so on. With growing recog-
nition of farms as real systems with their own
unique structure and function, it became
increasingly clear that off-farm employment
and the complex objectives of farm families are
as important as agronomic considerations.
Social sciences and biophysical sciences were
forced to develop a common conceptual frame-
work in order to work together.

Regional systems
While the evolution towards farming systems
research brought together economists, agrono-
mists and animal scientists, the interest in
regional systems brought in other groups of bio-
physical scientists, such as silviculturists and
hydrologists, and another group of social scien-
tists, such as sociologists, geographers, anthro-
pologists and those interested in community
development. The farming systems and regional
systems specialists had 'discovered' each other.
The development of a conceptual frame-
work acceptable to these various disciplines at
the regional level continues to be the chal-
lenge for the 1990s. It is complicated by the
fact that teams are trying to integrate expand-
ing system performance criteria and an
expanding population of target beneficiaries.
Further, the institutional arrangements and
information management systems are not yet
in place to support watershed-level systems

3.1.3 Expanding system performance criteria
Alongside the expanding scale of target sys-
tems, FSR teams have been expanding the crite-
ria used to evaluate the performance of the
systems they are developing, moving from a
predominance of system productivity towards a
balancing of multiple criteria that include crite-
ria factors associated with stability and sys-
tem/resource interactions. I have subdivided
this evolution into productivity, stability, and
sustainability (see Fig. 3.1.2).

The primary challenge for agronomists in the
1960s and early 1970s was to find a way to
measure system yield, rather than individual
crop yield. Two important breakthroughs were
the development of the land equivalent ratio
(the amount of land to be planted in monocul-
ture plots to equal the production obtained from
a polycultural system) and the land-time equiv-
alent ratio that looked at production over an
equal time period as well as area. In the 1970s
cropping systems teams saw that system pro-
ductivity should be measured in relation to the
system's primary limiting factors and that these
were not always land area and time7. Yields

FSR Understanding Farming Systems

began to be reported in kg mm- of rainfall or in
tons person-day-1 of labour. It began to be clear
that farmers sometimes use different criteria to
evaluate different crops within the same sys-
tem. In many cases farmers evaluate one crop
in terms of yield area-1 while another is evalu-
ated in terms of yield volume- of seed planted.
On the whole, these new perspectives on pro-
ductivity are still seen as challenges by the
research establishment.

Resource-poor farmers have no choice but to
consider productivity and risk simultaneously.
The risks that influence their decisions are
both ecological and economic, including
unpredictable market prices and the insecurity
caused by unclear land tenure. Short-term
family survival obviously must take prece-
dence over potential long-term economic
returns. Resource-poor farmers have no
choice but to design crop and livestock systems
that trade-off higher productivity for reduced
risk, and those pushed into fragile environ-
ments with poor soils and unpredictable rain-
fall are particularly vulnerable.
Risk is easy to recognize, but not easily
incorporated as a research criteria. For most
research teams, the simplest proxy for risk was
variability in yield or in efficiency. But on-farm
research was only just starting and many years
of data are needed to measure year-to-year
variability. Many FSR teams resorted to using
spatial variability as a proxy for temporal vari-

1970s 1980
productivity product

O/i O/1 [
outputs (0)

system syst

inputs (I)

ability, measuring system production along a
gradient of soil moisture, for example, as a
proxy for the year-to-year variability in rainfall.
Gordon Conway, in 19858, made a distinction
between what he called system stability (the
degree to which production is constant in the
face of small disturbances over time) and system
sustainability (the ability of a system to maintain
productivity in spite of major stress). Although
his definition of sustainability was not commonly
accepted, he made a very important contribution
by suggesting that stability, sustainability, equity
and productivity are all important system perfor-
mance characteristics. In the 1980s FSR teams
began routinely to apply system performance cri-
teria in addition to productivity.

Sustainability has become one of those terms
that will probably need to be abandoned
because of its multiple meanings to different
people. Most people who use the term are sug-
gesting that long-term environmental costs
need to be taken into consideration when evalu-
ating short-term economic benefits. FSR is cur-
rently struggling with ways to operationalize
this concept.
At least three different schools of thought on
how to use the concept are evolving: some view
sustainability as synonymous with a group of
technologies such as the use of animal manure
and integrated pest management viewed as
being more 'ecological'. Others define it as a
characteristic of the relationship between a

s 1990s
ivity productivity
S stability

/- socioeconomic
ie resources

em system


Fig. 3.1.2. FSR's expanding system performance criteria. FSR's conceptual framework has expanded from an
initial emphasis on system productivity in the 1970s, to an emphasis on both productivity and stability in
the 1980s, to an emphasis on a complex set of criteria related to sustainability in the 1990s.

46 Chapter 3

system and the resources upon which the sys-
tem depends, emphasizing the need to avoid
environmental degradation. A third group view
sustainability as a measure of intergenerational
equity. All have an equal claim to the legitimacy
of their definitions, but the three approaches
have different implications for FSR's conceptual
framework. The first will have a minimum
impact on the FSR framework as the same
technology can be appropriate in one system
and completely unsustainable in another. The
second use means that natural resources should
be included explicitly within the FSR conceptual
framework. The third use requires that future
generations be included within the FSR target
Clearly we cannot wait 50 years to decide if
a system is sustainable, as the second or third
approach would require. Possible indicators of
'sustainability' are currently being explored by
many different institutions and these include:

* Measuring all inputs and outputs and calcu-
lating changes in system efficiency (the total
factor productivity approach).
* Monitoring indicators of natural resource
productivity and modelling the probable
future impact of these changes, such as soil
* Setting up benchmark sites where many fac-
tors can be measured, in order to identify
minimum data sets (indicators) for use by
researchers working in similar environments.

1970s 1980s
small farmers men and

3.1.4 Expanding target beneficiaries
While FSR was expanding the systems' scales
and system performance criteria, it was also
expanding and differentiating its target benefi-
ciaries. Taking a systems approach to agricul-
tural research makes sense regardless of farm
size or farmer access to resources, but many peo-
ple view cropping and farming systems research
as synonymous with working with resource-
poor farmers. This confusion occurred because
in many countries the decision to adopt a sys-
tems approach in agricultural research and the
decision to work with poor farmers occurred at
roughly the same time. And, of course, it is eas-
ier to argue the merits of taking a systems
approach when the target systems are more
complex and less well understood.
Equity-related policy issues have always
played a significant role in the evolution of FSR.
In the 19 70s when on-farm research began, most
teams assumed that the primary beneficiaries of
their efforts would be the farmers in the commu-
nities where the research was conducted. In the
1980s equity related policy again affected the
evolution of FSR when gender issues were intro-
duced and women became explicit target benefi-
ciaries. In the 1990s the issues of sustainability
and intergeneration equity added future genera-
tions to the list of target beneficiaries of FSR.
I have subdivided the phases of expanding
beneficiaries into: small farmers, women and
urban poor, and future generations, as depicted
in Fig. 3.1.3.

men and women
future generations

t 9* m

Project Project
Benefits Benefits


Fig. 3.1.3. FSR's expanding target beneficiaries. FSR's conceptual framework has expanded from an initial
emphasis on small farmers as the beneficiaries in the 1970s, to an emphasis on both small farmers and
women in the 1980s, to an emphasis on small farmers, women and the next generation in the 1990s.

FSR Understanding Farming Systems 47

Small farmers
In the 19 70s one of the easiest ways to start an
argument when FSR specialists got together
was to ask someone to define what they meant
by a 'small farmer'. What was clear to most FSR
teams was that they were trying to develop bet-
ter technologies for farmers with fewer
resources. The selection of what type of
resource-poor farmer to work with differed
across different FSR teams. Institutions with a
commodity focus, such as the Consultative
Group on International Agricultural Research
(CGIAR) centres, worked with national agricul-
tural institutions in geographical regions where
their commodity was important or where the
environment meant that their commodity had
high potential. In many countries, national pol-
icy directed this effort towards regions with
resource-poor farmers. Cropping systems and
farming systems were selected for experimenta-
tion and farmers in the region operating similar
systems or recommendation domains were
assumed to be the primary future beneficiaries.
Many of the FSR projects financed by exter-
nal donors were often designed with a bias
towards resource-poor farmers in general,
rather than towards a particular crop or live-
stock enterprise, or, indeed, a particular farm-
ing system. In theory these teams had the
option of deciding which crop or animal sys-
tem, or non-agricultural household enterprise
to work with, based on an analysis of the
potential benefits to their target population.
Two main problems made this difficult in prac-
tice: on-farm research often began after a very
superficial appraisal of the local situation, and
the biases of team members greatly affected
the selection of the experimental focus. Too
often, the intended beneficiaries gained little
from the project.

During the 'small farmer' phase described
above, FSR teams assumed that local farm fami-
lies were their beneficiaries. Most of the scien-
tists were men and, particularly in Latin
America, Africa and Asia, they tended to inter-
act primarily with male farmers. In the late
1970s and early 1980s, as target systems
expanded and teams began to look at farm and
community systems, the involvement of
anthropologists and sociologists grew. The

issues of gender, equity and specific benefits to
women began to be explicitly addressed.
Gender was incorporated into the FSR con-
ceptual framework in different ways by different
institutions. Many simply changed the way they
wrote up their research proposals in order to
increase their chances of getting money from
donors that made gender an 'issue' and contin-
ued with their old approach. Other institutions
began to include gender as a variable as they
measured labour inputs and the flows of bene-
fits. And some adopted an equity approach and
set up special programmes targeted directly to
Regardless of motivation, the incorporation
of gender issues led most FSR teams in the
1980s to subdivide 'the black box' usually
labelled 'family' or 'household' in their farm sys-
tems diagrams into two categories: males and
females. Many FSR teams began to recognize
that gender differentiation meant more than just
simply measuring which gender provides the
most labour into a particular system. It also
involves an analysis of who makes key decisions
on resource allocations and marketing, and who
receives and controls the money when com-
modities are sold. While it has not been easy to
channel the benefits of FSR towards women
(this is not surprising given that equity involves
shifts in power) there is little doubt that changes
in the FSR conceptual framework to incorporate
gender have led to a better understanding of
how farm systems work.

The next generation
In the 1980s the beneficiaries of FSR were subdi-
vided by gender. In the 1990s they were further
subdivided into present and future generations.
The design of alternative systems and the
development of appropriate technologies now
becomes a question of how to predict benefit
streams to different potential beneficiaries.
One interesting example is the analysis of
the Plan Sierra Project in the Dominican
Republic carried out by Alain de Janvry9. In a
report written for the International Fund for
Agricultural Development (IFAD), de Janvry
and his colleagues analysed data from the pro-
ject, suggesting that intergenerational equity
could be a good way to measure sustainability.
He conducted an appraisal of the development
project from two perspectives: one from the

48 Chapter 3

standpoint of the present generation at one
point in time and one from the standpoint of
the next generation at another point in time.
This differed from the typical approach by
including externalities or charging a project for
the depreciation of natural resource stocks.
Basically, de Janvry's approach requires the
actions of the present generation to maintain
resources for use by the next generation.
This approach requires at least 40 years of
data, assuming 20 years for each generation.
While this is seldom available, approaches can be
developed that build on the concept. Since the
bridge between generations is the maintenance
of the productive potential of natural resources,
changes in resource productivity can, perhaps,
be used as a proxy for intergenerational equity.
One thing is very clear: the concept of time must
be more explicitly taken into consideration along
with measurements of the changing productive
potential of natural resources.

3.1.5 Operational implications

The expansion of FSR's conceptual framework
has made it almost impossible for a single insti-
tution to implement FSR. The expansion in sys-
tem scale, performance criteria and target
beneficiaries demanding the involvement of so
many disciplines over such a long time period,
makes it highly unlikely that one institution
could organize and manage an FSR initiative
using the expanded conceptual framework. The
involvement of multiple disciplines from multi-
ple institutions using multiple performance cri-
teria with benefits flowing to multiple
beneficiaries, makes FSR implementation even
more complex.
However, the common sense sequence in
FSR from analysis to design, evaluation, and
dissemination stages, has not changed with the
expansion in the conceptual framework.
To implement FSR within the expanded con-
ceptual framework it is obvious that different
types of institutional arrangements will be neces-
sary. It is equally obvious that the involvement of
multiple institutions and larger teams with repre-
sentatives from more disciplines will require more
efficient information management processes.

Institutional arrangements
All the following types of institutions have a
role to play:

* Farmers and farmer organizations.
* Community level institutions such as munic-
ipalities and marketing associations.
* Watershed-level institutions such as irriga-
tion management organizations and forest
management associations.
* Research organizations involved in agro-
nomic, soil, livestock and forestry research.
* Extension research organizations involved in
agronomic, soil, livestock and forestry exten-
* National, state or provincial institutions
involved in policy decisions affecting the
* International donor agencies.
What type of institutional arrangements need
to be developed? Assuming that various institu-
tions have indicated that they share a common
goal and that a minimum of financial resources
are available, the institutional arrangement
must provide:
* Representative governance.
* Efficient administration.
* Effective programmes (Fig. 3.1.4).
All participating institutions must have con-
fidence in the governing body. The administra-
tive structure must ensure that programmes are
following the policy guidelines of the governing
body, making efficient use of financial
resources. Programmes must integrate the dis-
ciplinary expertise of all participating team
members to ensure that resource use meets the
criteria defined by the governing body.
These institutional arrangements can be
called consortia, coalitions, alliances, associa-
tions or networks. The name does not matter.
What is important is that they develop represen-
tative governance, efficient administration and
effective programmes. All three of these areas are
a significant challenge, but, for some reason, the
question of how to develop representative gover-
nance has been a particular problem. This seems
to stem from the fact that donors providing finan-
cial support find it easier to ask larger interna-
tional or regional institutions to manage the
funds. They find it difficult to release the funds to
institutions over which they have no control.

Information management processes
One lesson learned and re-learned by FSR teams
is that the product or output of FSR cannot be

FSR Understanding Farming Systems 49

summarized in concise recipes or static techno-
logical packages. FSR is a process that has no
neat beginning and end.
The agricultural systems found in any given
watershed or community are managed by farm-
ers, community leaders, directors of watershed-
level organizations, policy makers and so on.
Rather than defining the operational goal of FSR
as the development of new systems, perhaps the
objective of FSR should be operationally defined
as to provide decision makers with good infor-
mation and the knowledge necessary to use this
information to make better decisions (Fig.
3.1.5). This information and knowledge is what
local people demand from FSR teams. The chal-
lenge is to respond to this demand.
Cooperative research, and in particular
interdisciplinary research, is impossible without
an information management system that cap-
tures, archives, processes and disseminates
information. FSR's analysis, design and evalua-
tion cycles capture information from multiple
sources. This information is used by individuals
from different disciplinary backgrounds to con-
duct ex ante design and ex post evaluation of
alternative resource use systems. But these
internal processes are of no practical value
unless they respond to a real demand for infor-

Target systems




Types of

mation and if they do not lead to a dissemina-
tion of information to farm, community, region
and national decision makers.
As in the case of new institutional arrange-
ments, the development of efficient information
management processes to support FSR has only
recently been recognized as of critical impor-
tance. Most FSR initiatives use their computers
to analyse their field-level data, to write reports
and technical papers and to develop their pro-
ject budgets. Very few are developing databases
and using them to capture information from
multiple sources, systematically evaluating
alternatives using multiple criteria, or systemat-
ically transferring information to multiple
potential beneficiaries.

3.1.6 Concluding remarks

Over the years I have had many opportunities to
discuss these issues with colleagues from past
projects as well as with other FSR practitioners,
and many of the ideas that I have outlined
probably originated with them.
An interesting question, of course, is, what
will happen next? I have mentioned the problems
facing institutions trying to operationalize the


Fig. 3.1.4. FSR's institutional arrangements. In order to implement FSR using an expanded conceptual frame-
work, institutional arrangements that include a broad range of stakeholder institutions must be developed.

50 Chapter 3

Target systems

Research and


Decision makers

Fig. 3.1.5. FSR's information management. In order to implement FSR using an expanded conceptual frame-
work, information management processes must be developed that link local decision makers and FSR
implementation processes.

complex framework that has evolved. Many con-
sortia and networks are now being formed and
valuable lessons will be learned from their experi-
ences. A major problem (in particular in this post
cold-war era where money for agricultural devel-
opment is scarce) is how to set up and manage
the complex institutional arrangements that are
a prerequisite for FSR implemented within its
expanded conceptual framework.
I feel strongly that the key to making these
complex institutional arrangements operational

is to develop more efficient information man-
agement processes linking all participating
partners and stakeholders. While the challenge
is great, there is no reason to think that the next
generation of FSR practitioners will not be cre-
ative enough to find a way to make things work.
Fortunately the information revolution has
arrived at exactly the right time to help create
new information management tools. The next
10 years in the evolution of FSR should be very

1. Norman, D.W., 1968. Why practice intercropping? Samaru Agricultural Newsletter. 10 (6). 107-16.
2. Margalef, R., 1968. Perspectives in Ecological Theory. University of Chicago Press, Chicago.
3. Bradfield, R., 1972. Maximizing food production through multiple cropping systems centred on rice. In:
Rice, Science, and Man. International Rice Research Institute. Los Bafios, pp. 143-63.
4. Conway, G.R., 1985. Agroecosystem analysis. Agricultural Administration, 20, 31-5.
5. Fresco, L.O. & E. Westphal, 1988. Hierarchical classification of farming systems. Experimental
Agriculture, 24, 399-419.
6. Hart, R.D., 1982. An ecological systems conceptual framework for agricultural research and develop-
ment. In: Shaner. W.W., P. Philipp & W.R. Schmehl (Eds) Readings in Farming Systems Research and
Development. Westview Press, Boulder, Colorado, pp. 44-58.

FSR Understanding Farming Systems

7. Zandstra, H.G., E.C. Price, J.A. Litsinger & R.A. Morris, 1981. A Methodology for On-Farm Cropping
Systems Research. The International Rice Research Institute, Los Bafios.
8. Conway, G.R., 1985. (Op. cit.).
9. de Janvry, A., E. Sadoulet & B. Santos, 1994. Project Appraisal for Sustainable Rural Development:
Notes for IFAD's Operational Guidelines. International Fund for Agricultural Development, Rome.

Mike Collinson

The growing recognition of the link between poverty and environmental degradation is forcing a much-
needed reconciliation between the traditional physically based definitions of zones and people based defini-
tions. The farming system can be an effective interface between these two traditions and form a basic unit
for agricultural research and development.

3.2.1 Introduction

Historically, useful typologies have been partic-
ular to purpose. As a means of stratification
they reduce sources of variation and increase
cost/effectiveness in sampling and extrapolation
within a broad universe in the case of most
FSR the resource-poor farmers of the world.
Two streams of activity have dominated typolo-
gies used in research and development for
small-farmer agriculture. At the macro level,
biophysically based agroclimatic and agro-
ecological zoning and, at the micro level, farm
classification. Agroecological zoning has been
developed by land management professionals,
primarily to classify land according to
potential1, whereas farm classification is a tool
of the farm management profession, its primary
purpose to compare the performance of
managers operating similar farms2.
The growing recognition of the link between
poverty and environmental degradation is forc-
ing a much-needed reconciliation between the
traditional physically based definitions of zones
and people-based definitions. The farming sys-
tem can be an effective interface between these
two traditions and form a basic unit for agricul-
tural research and development.

3.2.2 Agroecological zoning
FAO's pioneering work
In the late 1970s and early 1980s, the Food
and Agriculture Organization of the United
Nations (FAO) completed agroecological zoning
for the major regions of the world based on cli-
mate, soil and terrain criteria. The main pur-
pose was to measure agricultural potential and

define optimal land use. In one innovative appli-
cation, zoning was used to show which coun-
tries of the world should be able to feed
themselves into the future3. The productive
potential of the zones was related to population
and its growth, estimates of carrying capacity
and, beyond that, calculations of future self-suf-
ficiency at the country level. For the first time
countries with long-term problems of food self-
sufficiency were identified through a logical,
analytical process.

Issues in the use of biophysical based
Biophysically based zoning has been widely
used, and sometimes abused, as a policy plan-
ning tool for agricultural research and develop-
ment4. Planners have gone so far as to designate
the zones of highest physical potential as the
only areas where growers are serviced by public
marketing, extension and credit institutions.
Examples in eastern Africa showed areas
denominated as 'very suitable' for sorghum
which are, on the ground, dominated by maize,
the staple food chosen by local farmers.
Zoning has played, and continues to play, an
important role in agricultural research for three
key reasons:
* To identify, for plant breeders, sets of condi-
tions with common climatic constraints
within which they must work, and provide a
framework for testing and comparing mate-
* To identify relatively homogeneous sets of
climate and soil characteristics that can be
dealt with by one applied research pro-
gramme in natural resource management.

Chapter 3

* To show where else applied research results
will be relevant, and facilitate transfers
between geographically separate but rela-
tively homogeneous sets of climate and soil

These roles also create distortions because
they don't allow for farmers' decisions.
Agroecological zones offer a range of cropping
opportunities to farmers and, at the same time,
within each crop opportunity offer a set of
possibilities; for example, maize is an oppor-
tunity, but the farmer may choose a maize
which matures in 110-200 days. Farmers'
choices from these 'within crop' sets are deter-
mined by their socioeconomic circumstances
and articulated in their farming system.
Preoccupation with yield has kept applied
research firmly focused on the upper boundary
of the 'within crop' possibility set. For example,
in an agroecological zone characterized by a 6-
month rainfall season, a 180-day maize offer-
ing the greatest yield potential will be the
logical choice for conventional breeding pro-
grammes. Clearly, maize with maturity periods
below 180 days are cropping possibilities for the
zone but applied research will pay little or no
attention to them, even though they will often
be the preferred choice of farmers, for at least
three possible reasons:

* Low power resources may stop farmers
opening the land until it is thoroughly wet-
ted. The earliest feasible planting time may
be 20 days after the start of the rains, sug-
gesting a 160-day variety.
* Local maize prices may be 300-400%
higher several weeks before the main har-
vest, suggesting a maturity period of around
145 days to exploit this profitable market.
* Farmers may give priority to planting
another crop before or after the maize, sug-
gesting a maturity period of 110-120 days.

Varieties and management practices identified
as 'optimal' will often be totally irrelevant to the
choices made by farmers as a result of local cir-
cumstances. Indeed, the ranking of crops and
crop varieties on the basis of physical potential
for an area can be stood on its head when farm-
ers' own criteria are used, rather than those of

3.2.3 FSR and farm classification

Recommendation domains
As the Farming Systems Support Project of the
University of Florida said in 19876: At one
extreme, we do not have sufficient resources to
carry out a specific research program for every
individual farmer. At the other extreme, it does
not make sense to try to develop a single
research agenda relevant to all farmers in a
country. We must compromise between these
two extremes and plan research relevant to
groups of farmers'.
The tradition of grouping farms and farm-
ers, already acknowledged in Anglophone
countries and in France, was carried into devel-
oping-country agriculture in the 1970s. Upton,
in 19737, for example, focused on the need for a
process dealing with groups rather than indi-
vidual farmers. One early struggle was the move
away from survey work based on administrative
areas, which confounded understanding of
farmers' decision-making by averaging data
across more than one farming system8. The
work took a major step forward in the mid
1970s when the International Maize and
Wheat Improvement Center (CIMMYT) devel-
oped the concept of the 'recommendation
domain' (RD); a group of farmers operating the
same system and for whom the same new tech-
nologies would be appropriate. The RD concept
recognized spatial and hierarchical dimensions,
with hierarchical variation between systems
arising from differences in farmer resource
Criticism arose from two perceptions. First
that manpower-intensive local surveys are a
prerequisite to system differentiation and are
impractical over wide areas. Second, that rapid
rural appraisal could provide only qualitative
information and this was unattractive to agri-
cultural economists trained in a quantitative
Collinson, writing in 19969., held that varia-
tion in six key parameter sets climate, soils.
topography, culture, market opportunity and
population pressure cause spatial diversity
across farming systems. However, he saw the
ideal minimum data set to identify RDs as hav-
ing three groups of criteria to capture the
effects of both spatial and resource endowment

FSR Understanding Farming Systems 53

* The spatial pattern and scale of farm activities.
* The practices used to manage these activities.
* The calendar of the application of these
practices over the year.
These three groups feature empirical informa-
tion about the systems operated by farmers. The
data required can be captured by low cost rapid
rural appraisal methods where up-to-date sec-
ondary data is not available.

Issues on RDs
The strength of the RD concept is its firm focus
on grouping farmers the managers who will
take the decisions on using new technologies.
Aspects under continuing discussion include the
acceptable level of compromise in reducing
sources of variation between farms, and the
issue of domain boundaries. Viewed more
widely, there are clear shortcomings to the RD
concept in the search for a conceptual frame-
work for agricultural research and development.
Byerlee, in 198710, distinguished three types
of variation within farming systems; between
farms, within farms and between years. He
noted: A point is soon reached where between-
farm variability is less important than variabil-
ity within farms and across years and the
benefits of [further] sub-grouping rapidly
diminish'. The acceptability of the compromise
depends on the importance of sources of
between-farm variation that are not captured
by the criteria used in system definition.
Fortunately much of this variability is caused
by two sources which do not create 'between-
system' variability: age and its effect on motiva-
tion is one, and differences in managerial
competence or natural ability as a farmer is the
other. Both cause between-farm variability but
are sources common to all systems and both are
particularly useful in identifying subgroups
within farming systems on which scarce
resources for research and development
resources might be focused.
A second area of continuing discussion,
which applies to all zoning, is that of system
boundaries and the extrapolation of results.
Perrot and Landais addressed the issue in
199311, saying: All segregative methods pre-
sent the disadvantage of focusing too much
attention on the limits separating each type, on
the definition of type content and therefore of

their specificity which is subordinated to the
limits between them. In other words, the bor-
ders are all important and the centre is
neglected, when the contrary would be more
consistent with the constructivist approach'.
Just as with climate and market opportuni-
ties, one farming system usually merges into
another. Where there is a stark change in a
causal factor, a dramatic change in topography
for example, a boundary may be clearly identifi-
able, more usually there is a graded transition
and no clear boundary. It is a phenomenon
which can be managed locally. Research and
development activities need to target the clearly
defined populations within domain boundaries,
and programmes, such as farmer-led experi-
mentation, need to be located unambiguously
within farming systems. Social forces and the
market will stimulate diffusion to the fringes
and thus define where new technologies and
policies lose their relevance1.
The RD concept has, like zoning, sometimes
threatened to overreach itself. Even early docu-
mentation from CIMMYT13 tentatively suggests
that a farmer may be a member of two
domains. Similarly, the biophysical importance
of differentiating landscape types provides the
temptation to be all things to all people14 but
breaks the ground rule: one farmer, one
domain. The original RD concept based on
grouping farmers as decision makers has been
further compromised by the application of the
term 'RD' to a technology. Not only is it confus-
ing to have two contrasting definitions, but the
technology-based definition wholly ignores the
human side of the adoption equation, which
FSR has struggled to bring into full partnership
with the biophysical.

Shortcomings in the RD concept
The evolution of conceptual frameworks shows
how purpose in FSR has continued to widen.
Three factors, in particular, have added to the
complexity of this process:
* The dynamics of farming systems in res-
ponding to widening market opportunities
and increasing population densities1 5.
* The reconciliation of the goals of individual
farmers, their community and society at
large, particularly with respect to environ-
mental sustainability.

54 Chapter 3

* The strong interactions between technology
adoption and policy manipulation16.
These same factors have compounded the com-
plexity required in typologies as a framework for
agricultural research and development. The
recommendation domain concept has not
evolved to embrace them, but new concepts and
techniques are carrying the challenge forward.

3.2.4 Recent developments and current status
1986 saw a landmark meeting on agroecological
characterization17. Though many contributors at
the meeting emphasized the importance of
socioeconomic parameters in a framework for
agricultural R & D, participants concluded: 'It was
not agreed that the time was ripe to incorporate
socioeconomic parameters into databases along
with data on other environmental attributes.'
Contributions included pioneering studies that
combined the use of macro-level geographical
information systems (GIS) with socioeconomic
data from national census results18, and from
rapid rural appraisal methods19.

The sustainability issue
During the late 1980s and early 1990s, the call
for a balance in biophysical and socioeconomic
considerations was strengthened by a growing
awareness of the sustainability issue. The
Brundtland Report in 1987 provided a strong,
early stimulus and was followed by the UN
Conference on Environment and Development in
1992. In 1989 the Technical Advisory
Committee recommended a restructuring of the
Consultative Group for International Agricultural
Research (CGIAR) to bring more of its science to
bear on natural resource management research.
The basis of the restructuring was the acceptance
that, beyond natural processes, human decisions
at levels from the farm to the globe are primary
causes of resource degradation. The shift in
CGIAR programming brought new impetus to the
search for balance between biophysical and
socioeconomic dimensions of the resource man-
agement problem. Similarly, as the UN Task
Manager for the implementation of Chapter 10 of
Agenda 21, FAO moved away from its historical
emphasis on climate and soils and towards a
holistic people-based approach to the planning
and management of land resources. The sustain-

ability issue drew both geographers and ecologists
into the widening disciplinary mix of the emerg-
ing paradigm for agricultural research and devel-

Growth dynamics
Recent work has begun to erode the static nature
of typologies. Research by Weber, Smith and
Manyong20 identifies four broad types of agricul-
tural systems (named as 'research domains') -
market intensive, market extensive, population
intensive and population extensive designating
market access and population increase as two
key drivers in farming systems. The authors per-
haps go too far in claiming 'the most relevant
information about a system and its sustainability
is related to its evolutionary pathway rather than
its current characteristics'. It is true, however,
that much of the historical rejection of research-
derived technology has been a failure to under-
stand labour and cash as greater constraints
than land on farmer livelihoods.

Systems hierarchies
Fresco, in 199521, made the important point: 'If
we want to explain, rather than solely describe.
a phenomenon, processes at both higher and
lower scales must be studied also'. Systems hier-
archies had already been recognized as valuable
tools, for example in agroecosystems analysis in
the 1980s22.
In addition, farmers were increasingly recog-
nized as de facto land managers. Both were
accompanied by a strengthening perception of
the need for linkages between farmers' actions,
national environmental policies and global con-
ventions important factors highlighting the
relevance of a hierarchical framework. In the
example framework at Table 3.2.1 human deci-
sions (farm, community, local, regional,
national policy, international convention) influ-
ence activities at several levels in both economic
and ecological hierarchies. The impact of their
decisions not only has repercussions on the
adjacent levels within the hierarchy but also,
importantly, interacts across the two hierar-
chies. So economic decisions have ecological
implications and vice versa.
Operationally the 'decision point' identifies
institutional levels which offer leverage on both
the economic and ecological hierarchies. The
key points on the decision hierarchy are farmers

FSR Understanding Farming Systems

Table 3.2.1. Three hierarchies: a framework for agricultural R & D.



Agro climate
Agro ecology

Agroecological zone

Major topographical
features (river basin,
forests, mountains)


Resource management domain

Recommendation domain
Farm enterprise

Landscape unit
Farm system

Resource niches/fields

Village authority

Farm family
Family members

and policy makers. With effective institutions
open to the market, these decisions help shape
the production environment in which farmers
make their decisions. With weak or 'captured'
institutions, as with market distortions, policy
decisions have less influence. Farmer and com-
munity decisions on change directly impact the
landscape and farm levels. As more farmers and
communities adopt the changes, their accumu-
lating decisions have repercussions at higher
levels23. Linking policy and farmers' decisions
in the same framework embraces the interac-
tion between policy formulation and technolog-
ical innovation.
The surge of interest in sustainability has
stimulated the notion of resource management
domains (RMDs). These are variously defined
and some retain the early emphasis on land
potential. One more general and more flexible
definition comes from Esawaran writing in
199624: 'Resource management domains are
landscape units delineated on the basis of simi-
larity with respect to response to management'.
Esawaran located these in a hierarchical frame-
work at an intermediate level between agro-

economic zones (AEZs) and RDs. He describes
them as a set of biophysical parameters bound-
ing a group of technical problems with causes
peculiar to their characteristics and a parallel
set of technical solutions. The farm system or
RD is shown as common to both the economic
and the ecological hierarchies, highlighting the
fact that farmers' decisions and actions are a
dominant influence on both hierarchies.

3.2.5 The current state of the arts

As Hart has pointed out, system hierarchies
offer a valuable conceptual and operational
framework for agricultural research and devel-
opment. In addition, GIS and remote sensing
(RS) techniques both make a strong contribu-
tion to useful application by allowing the over-
laying of spatially referenced data sets at
different hierarchical levels. Users may draw on
those data sets needed for their particular appli-
cation. This flexibility allows use of the data sets
for a variety of purposes making the framework
more general and less 'purpose specific' than
older typologies. However, the historical priority

Decision point
conventions and organizations

Country groupings
National policy
National institutions
Regional institutions

Local government
NGOs, farmer




56 Chapter 3

given to database compilation in climate and
soils limits the scope of current GIS applications
and the challenge of accumulating new data-
bases is a critical contemporary constraint.
One way to apply this type of hierarchical
framework in agricultural research and devel-
opment, particularly in the sustainability con-
text, is to use Naisbett's 1994 line: 'act locally,
think globally'25. It is widely accepted that
resource degradation begins locally. It is
unnecessary to research all levels of the hierar-
chies. The key is effective local diagnosis to
identify the decision points driving local activi-
ties, particularly those causing degradation.
The issue of 'full' or 'partial and focused' sys-
tems analysis, and the use of qualitative or
quantitative data have been controversial since
the early days of FSR. CATIE, for example,
sought full quantification and considered all
the parameters of the farming system as vari-
able. CIMMYT26, on the other hand, applied
low-cost techniques providing largely qualita-
tive information to gain an understanding of
the wider system. Carrying a 'partial and
focused' approach into this wider framework
would identify the hierarchical levels and deci-
sion makers, many of which will be external to
the farming system, yet crucial to changing the
activities causing local degradation27.
A hierarchical framework embracing ecol-
ogy, economics and institutions has important
implications for the organization of institu-
tions and their coordination in its effective
application. Tims in 199528 argued: 'Research
should increasingly centre around studies of
farm households, each with farming as one of
its activities, but among a number of other
options to use its scarce resources, and with
trade-offs which affect the choices in the field
of agriculture. Also account must be taken of
the character of production decisions in a
number of cases as derived decisions. ...
Households interact with policies and markets
and research cannot truly answer policy ques-
tions unless it traces those market relations,
with price and income formation and the
responses to those by households and by gov-
ernments'. Tims' statement highlights the
range of external influences on farmers' deci-
sions. It points to an understanding of farmer
decisions as a prerequisite to progress in agri-
cultural development.

The farming systems movement is in the
forefront of equipping professionals with that
understanding and the wider hierarchical
framework offers a more robust context for
their activity. Permanent field teams are justi-
fied by the need to understand small farm
households and apply that understanding at
several levels. With their eyes on resource
degradation Carter and others have said
(rightly in my view): 'put simply how can we
predict what we don't understand?'29

3.2.6 A scenario for the future
To mobilize current knowledge in an integrated
way requires a combination of a wider concep-
tual framework, institutional change as well as
improved information management tech-
niques. Each institution, and indeed each indi-
vidual, needs to know where they are, and how
they contribute, in a strong operational frame-
work. The ability to position themselves in this
framework tells them who their logical part-
ners and clients are, and how information
needs to flow if they are to be successful in
their role.
The improving resolution in remote satellite
sensing and progress in digitizing images,
means that the spatial pattern of agricultural
activities may offer a global basis for typing
farming systems with less need for field work.
But this should not, however, induce the search
for the 'holy grail'. Countries dependent on
their small-farm sector, with large numbers of
farmers and few professionals, will still need
local expertise30. FSR teams remain important
institutional innovations for many countries.
NGOs, farmer associations and publicly funded
agricultural services might be appropriate vehi-
cles for their operation, with a core task of par-
ticipatory technology identification.
Local FSR teams are the key to the effective
exploitation of both human and biophysical
information. They will understand how farmers
exploit their RMDs in the operation of their
systems. They will also understand which
enterprises are important sources of degrada-
tion and where. Such teams need access to the
widest range of technical findings on the
management of the RMDs with similar profiles,
particularly for those enterprises contributing
most to degradation.

FSR Understanding Farming Systems 57

As a long-term goal, local FSR teams should
be able to compare their local RD profiles with
other profiles from a global database, drawing
on those options that have succeeded under the
same conditions elsewhere. We are a long way
from this ideal. An ambitious first step will be to
identify RMDs and the technical and policy
opportunities associated with them. In the
short term, RMDs would be defined on the clas-
sic climate and soil criteria for which databases
already exist, supplemented by an accumulat-
ing database of technical research and policy
findings organized by RMDs. Local teams draw
on these for the RMDs in use by local farmers.
FSR teams will first identify technical and policy
opportunities for those RMDs in their local
space, and pre-screen these for their apparent
relevance to local RDs or farming systems. In
dialogue with local communities, those seen as
the most appropriate options are tested in a par-
ticipatory mode in farmers' fields.
At the same time, databases are built up for
RDs with input from the same FSR teams. These
would provide biophysical and socioeconomic
profiles of the sites where techniques and policy
measures had proved successful. Socioeconomic
profiles might include information on market
access, population pressure and 'calendars' for
the successful technologies. As such databases
gain recognition, their use by FSR teams would
widen and the process would evolve.
This route forward would have two short-
term benefits:
* It would diffuse the issue of incorporating
socioeconomic parameters into biophysical
* It would outflank the problem of weak data
on existing farming systems as an argument
against their adoption as primary develop-
ment units.
Once in situ, with research with farmers as their
core role, the mandate of local FSR teams could
be widened:

* Teams could help pursue greater system
understanding, including measurement of
key parameters for modelling of the dynam-
ics of the development path for example, or
the needs for sustainable farming, or the
sequencing of interventions.
* Teams can channel local information to lev-
els in the decision hierarchy where plans for,
for example tenure arrangements, com-
munity regulation structures, pricing of
water or remuneration for downstream costs
all influence local activities.
* Teams can contribute details of locally suc-
cessful practices in production and land
management to databases.

As the second of these supplementary func-
tions implies, the processes of social, as well
as technical innovation, can be facilitated by
local research31. Providing local FSR teams
with access to relevant global research
requires a major effort in database develop-
ment. Once such a database framework is
operational, local field teams themselves
would be well placed to enter, as well as to
use, information.
It is particularly encouraging that 'hard
core' disciplines are now reaching out towards
the human side of the development equation.
We have moved some way from Henry Nix's
assertion, made over 10 years ago, when he
said: 'I have argued that if it were possible to
predict the performance of any crop at any
location given a specified minimum set of site,
soil, crop, weather and management data, we
could indeed prescribe appropriate and relevant
technologies at the farm or even the field
level'32. Perhaps we will progress further down
the path if all institutions and professionals
with a role to play adopt a common hierarchical
framework, and if the will is found to adjust
both institutional mandates and structures to
play these roles in partnership.

1. FAO, 1978-80. Reports of the Agroecological Zones Project. World Soil Resources Report no. 48. FAO,
2. Blagburn, C.H., 1961. Farm Planning and Management. Longmans, London.
3. FAO, 1984. Land. Food and People. FAO Economic and Social Development Series no. 30. FAO, Rome.
4. Collinson, M.P., 1988. FSR in Evolution: Past and Future. Keynote address. 8th Annual Farming
Systems Research/Extension Symposium. Winrock International, University of Arkansas, USA.

58 Chapter 3

5. (Ibid.).
6. FSSP, 1987. Diagnosis in Farming Systems Research and Extension. Farming Systems Support Project,
Volume 1 FSR/E Training Units, Participant Manual. University of Florida, Gainesville, USA.
7. Upton, M., 1973. Farm Management in Africa: the Principles of Production and Planning. OUP. London.
8. Collinson, M.P., 1972. Farm Management in Peasant Agriculture. Praeger, New York.
9. Collinson, M.P., 1996. Social and economic considerations in developing Resource Management
Domains. Proceedings of the International Workshop on Resource Management Domains. August. Kuala
Lumpur, Malaysia. IBSRAM, Bangkok.
10. Byerlee, D., 1987. From Adaptive Research to Farmer Recommendations and Extension Advice.
Agricultural Administration and Extension, 27, 231-44.
11. Perrot, C. & E. Landais, 1993. Exploitations agricoles; pourquoi poursuivre la recherche sur les methods
typologiques? Les Cahiers de la Recherche-Developpements, 33, 13-23. Reprinted in English in Systems
Studies in Agricultural and Rural Development, J. Brossier, L. de Bonneval & E. Landais (Eds). INRA
Editions, INRA, Paris.
12. Collinson, M.P., 1996. (Op. cit.).
13. Byerlee, D., M. Collinson, et al., 1980. Planning technologies appropriate to farmers: concepts and pro-
cedures. Economics Program, CIMMYT, Mexico.
14. Fresco, L.O., 1995. Agroecological knowledge at different scales. In: Bouma. J. et al. (Eds) Ecoregional
Approaches for Sustainable Land Use and Food Production. Kluwer and CIP, The Netherlands.
15. Maxwell, S., 1986. Farming systems research: hitting a moving target. World Development. 14(1).
16. Baker, D., 1993. Inability of Farming Systems Research to deal with Agricultural Policy. Journal for
Farming Systems Research-Extension, 4(1), 67-86.
17. Bunting, A.H., Ed., 1987. Agricultural Environments, Characterisation. Classification and Mapping.
18. Scholz, U., 1983. The Natural Regions of Sumatra and their Agricultural Production Patterns a
Regional Analysis. Central Research Institute for Food Crops. Bogor. Indonesia.
19. Carter, S.E., 1987. Collecting and organizing data on the agro-socioeconomic environment if the cassava
crop a case study of a method. In: Bunting (Op. cit.).
20. Weber, G., J. Smith & M.V. Manyong, 1996. Systems dynamics and the definition of research domains
for the northern Guinea Savanna of West Africa. Agriculture, Ecosystems and Environment, 57, 133-48.
21. Fresco, L.O., 1995. (Op. cit.).
22. Conway, G.R., 1987. The properties of agroecosystems. Agricultural Systems, 24 (2), 95-117.
23. Collinson, M.P., 1996. (Op. cit.).
24. Esawaran, H., 1996. Biophysical considerations in developing resource management domains.
Proceedings of the International Workshop on Resource Management Domains. August. Kuala Lumpur,
Malaysia. IBSRAM, Bangkok.
25. Naisbett, J., 1994. Global Paradox. William Morrow. New York.
26. Byerlee, D., M. Collinson, et al., 1980. (Op. cit.).
27. Janssen, W.G., 1995. Aggregating economic knowledge for use at national, regional, sector and farm
level. In: Bouma, J. et al. (eds) Ecoregional Approaches for Sustainable Land Use and Food Production:
Proceedings of a Symposium on Ecoregional Approaches in Agricultural Research. Kluwer Academic
Publishers, Dordrecht, pp. 143-166.
28. Tims, W., 1995. Responses to Sections A & B. In: J. Bouma et al. (Eds) (Op. cit.).
29. Carter, S., P. Bradley, S. Franzel & J. Lynam, 1994. Spatial Variation as a Factor in Natural Resource
Management Research. International Institute for Environmental Technology and Management.
Stockholm Environmental Institute.
30. Delehanty, 1., 1993. Spatial projection of socioeconomic data using GIS: results from a Kenya study in
the strategic implementation of a livestock disease control intervention. In: Dvorak. K. (Ed.) Social Science
Research for Agricultural Technology Development: Spatial and Temporal Dimensions. CABI & IITA.
31. Carter, S., P. Bradley, S. Franzel & J. Lynam, 1994. (Op. cit.).
32. Nix, H., 1987. The role of crop modelling, minimum data sets and geographic information systems in the
transfer of agricultural technology. In: Bunting, A.H., 1987. (Op. cit.).

FSR Understanding Farming Systems 59

John Farrington

PRA has ... generated a sense of community ownership of development projects and processes, and a
recognition among administrators that farmer participation enhances the prospects of success. However, it
is increasingly being seen as a 'new orthodoxy', and, like all orthodoxy, it is attracting challenges of diverse
kinds. One has to do with intellectual property, viz. the argument that a number of the methods it
embraces predate the term 'PRA.

3.3.1 Introduction

A glance at the programmes for the interna-
tional FSR symposia held since 1981 reveals a
number of sea-shifts in major areas of concern,
many of them having implications for the devel-
opment and application of diagnostic methods.
The principal concern over the first five sym-
posia (1981-85) was to increase understanding
of small-farm systems, and to develop method-
ologies for diagnosis, implementation of
research and monitoring. By the seventh sym-
posium, concerns over the link between FSR,
macro-policy and communication had begun to
emerge, but the emphasis was still on 'how sys-
tems work'. By the eighth symposium, the
agenda had broadened to include major ses-
sions on gender and intra-household analysis,
farmer experimentation and natural resource
management. This set the tone for subsequent
meetings: by the 12th symposium in 1992
these new themes were joined by a major ses-
sion on different types of institution. The 13th
symposium in 1994 went further, considering
the role of agriculture in the generation of rural
This contribution looks back at these
changes and forwards to others which may
arise for diagnostic methods in the future.

3.3.2 Methods, policy context and
organizational change

This contribution argues that changes in meth-
ods have to be viewed in the context of major
changes in the mandates and structure of
research organizations, which, in turn, are
influenced by overarching policy imperatives.
The main interactions between policies, organi-
zations and methods are presented on a broad
canvas in Table 3.3.1. From the beginnings of
agriculture to the middle of the 19th century,
the imperative was to increase and secure

household food production. Farmer experimen-
tation was the sole means of achieving techni-
cal change in pursuit of this objective, and
informal interaction amongst farmers the
means of spreading technical change.
Agricultural science first became institution-
alized with the establishment of Rothamsted
Experiment Station in the UK in 1843, but
many scientists retained their roots in rural
communities, and farmer influence on the
research agenda was strong. There was little dif-
ficulty in diagnosing the priorities for research
in a 'whole farm' context. The first half of the
twentieth century saw two important shifts.
First, advances in plant and animal genetics
and in the understanding of pests and diseases,
plus the arrival of the internal combustion
engine and agrochemicals, opened the door to
the creation of specialist disciplines. Second, in
Europe this took place against a growing policy
imperative to increase the productivity of land
(this pressure was less intense in the USA, with
its lower population densities). Publicly funded
research expanded rapidly during this period,
but much of the expansion was in the form of
specialist institutes and departments to house
newly emerging disciplines. Their primary
approach was reductionist in the sense that
they assessed changes in productivity attribut-
able to specific technical changes by isolating
these from the remainder of the farming sys-
tem. In this context, priorities for research were
determined as much by the widening opportu-
nities offered by science as by careful diagnosis
of farmers' needs.
Policies to enhance the productivity of
export crops led the European colonial powers
to replicate these organizational structures and
research approaches in the south. With political
independence came a major drive towards
research on food crops. However, by this time,
many of the public sector research organizations
in developing countries had been structured

Table 3.3.1. The evolution of FSR diagnostic methods: policy context and organizational change.

1500 BC to 1850-1950 (S. Asia) 1950-1985 1985 to present
1850 AD 1850-1965 (Sub-Saharan Africa) 1965-1985

Policy Increase food
imperatives production *

and structure

among farmers

Methods/ Farmer
disciplines experimentation

Produce colonial export crops
Increase food production in north
Increase land productivity
especially in Europe

Shift from farm -- station -> laboratory-
based research
Shift towards commodity institutes
and commodity/discipline-based
departments in north; these
exported to south initially for export
crop production

Generalist researchers rooted in
rural communities gradually giving
way to specialist disciplines
Increasing reliance on reductionist
Research agenda increasingly driven
by science
Few comprehensive efforts at
diagnosis: to increase yields assumed
to be overriding objective
No social science skills deployed

* Increase food production in south
* Increase land productivity especially
in Europe

* Multidisciplinary teams, often led by
social scientists, formed within
national public sector research
institutes to diagnose farmers'
problems and opportunities and
prioritize research
* Efforts within national systems to
link field-oriented with disciplinary
researchers, and to create
research-extension links
* IARCs search for role in FSR, and
for appropriate working relationship
with national research services

'Increase yield' diagnosis correct in
high potential areas leading to success
of Green Revolution deriving from
reductionist approaches
Widespread failure of attempts to apply
similar diagnosis and approaches to
complex, diverse and risk-prone areas
To address above, economists (sub-
sequently other social scientists) conduct
extensive surveys to identify complexity,
yet rationality of farming systems in CDR
environments; 'rapid survey' techniques
towards end of period

* Reduce food surpluses in north
* Increase food production in south
* Do above in ways compatible with sustainable
natural resource management

* Reduced emphasis on diagnostic 'teams'; more
responsibility on individual scientists or groups of
scientists to identify clients' needs and design their
work in response
* Growing recognition that even with rapid methods,
public sector alone does not have the resources for
widescale diagnosis of the diverse needs arising from
agroecological and socioeconomic (including gender
and household) complexities
* Hence, growing emphasis on multi-agency approaches
and on having NGOs and farmers' groups diagnose
needs and create demands on public sector
* Efforts to increase responsiveness and accountability of
public sector enhanced by above, but also by
innovations in research funding, reward systems and
* Increasing reliance on rapid assessment and
participatory diagnostic methods
* These methods extended to understanding of
complementarities between on- and off-farm resource
management and their implications for research
* Increased efforts to understand implications of intra-
household decision taking and gender issues
Increased efforts to locate agriculture's contribution
within the multiple sources of livelihoods pursued by
low-income farmers
Growing interest in systems approaches to address
sustainability concerns in south and north

FSR Understanding Farming Systems 61

along discipline and commodity lines occa-
sionally even with separate specialist institutes
- in ways unrelated to the needs of small farm-
ers operating under complex, diverse and risk-
prone conditions. Researchers' conventional
diagnosis in this setting supported by only
limited contact with, often, the better endowed
farms was that the need to enhance yields per
hectare was paramount.
Research prioritized on this diagnosis gener-
ated widely adopted packages only where three
sets of conditions held: where pressure on land
was high, where production conditions were
favourable and stable, and where the main com-
ponents of farming systems could easily be
replicated on research stations. The rapid
spread of the Green Revolution in the irrigated
rice and wheat systems of South Asia is one
example. Over time, elements of such packages
were adopted selectively elsewhere as farmers
perceived them fitting into their systems', usu-
ally where extensive farming options had disap-
peared. Elsewhere, inadequate diagnosis of
farmers' requirements in a 'whole farm' context
led to low adoption levels, and to a widespread
view that farmers were 'backward' and unre-
sponsive to conventional economic stimuli
because of their unwillingness to adopt 'supe-
rior' technologies.
It is against such prejudice that the early
farm management investigations of Norman in
Nigeria2, Collinson3 and Ruthenberg4 in East
Africa, and Mellor5 in India have to be under-
stood. By examining how farm households
deployed their labour and other resources in
order to secure food requirements under high-
risk conditions, they gave a new dimension to
the understanding of farmer rationality. The
early review by Gilbert et al. in 19806 suggests
that much of this work was multidisciplinary,
but led by social scientists working with inten-
sive sample surveys of farmers' practices, fre-
quent recording techniques and questionnaires,
supplemented by the particular insights that
biological scientists were able to provide.
Bringing biological scientists and social scien-
tists together to investigate field realities in this
way had rarely been tried in small-farm agricul-
ture. It set the trend for farm surveys for the
1970s and early 1980s, and much of the early
institutionalization of FSR involved the forma-
tion of 'farm management' units or special

teams. It should be noted, however, that the tra-
ditions on which FSR drew were by no means
homogeneous: work in Senegal, for instance,
had led to the setting up of 'unites experimen-
Two exponents broke the methodological
mould of early FSR institutionalization:
Hildebrand8 by developing a 'sondeo' (literally
'sounding') method in Central America, which
many regarded as 'quick and clean', and
Collinson9 by experimenting with informal,
qualitative 'pre-surveys' using semi-structured
interviews and checklists. These had two pur-
poses: to characterize 'general attributes' and to
identify parameters vital to system improve-
ment for subsequent quantification in a formal
survey. Collinson also argued that, when con-
ducted in this setting, infrequent visits gave as
good data on critical parameters as frequent
visit methods. These were the precursors of
'Rapid Rural Appraisal'"0 and its derivatives.

3.3.3 Development of diagnostic methods in
the 1980s

Two further methodological developments were
reported in the early 1980s. First, the Rhoades
and Booth farmer-back-to-farmer model devel-
oped at the International Potato Center (CIP)
and described in 198211, stressed interdiscipli-
nary rather than multidisciplinary work. It was
characterized by interaction between farmers
and researchers in which the conventional pro-
ject cycle of diagnosis, experimentation, assess-
ment and dissemination can be replaced, for
instance, by approaches which begin with an
experiment and end with a survey. Practical
experience of interdisciplinary work at CIP led
to the following characterization of the main
stages of the model:

* In diagnosis the problem is identified jointly
by farmers and researchers.
* Interdisciplinary team research develops
potential solutions to the problem.
* Solutions are adapted to farmers' conditions
in on-farm testing.
* Farmers play a key role in evaluation and
further adaptation.

Second, the farmer-first-and-last model proposed
by Chambers and Ghildyal in 198512 entails

62 Chapter 3

'fundamental reversals of location and learn-
ing' and is characterized by:
* A diagnostic procedure involving learning
with farmers.
* Technology generation on-farm and with-
* Using the level of farmer adoption as a crite-
rion for evaluating research.
Chambers and Ghildyal outline the conditions -
including institutional conditions necessary
for the approach to succeed. These include
methodological flexibility and innovation, full
interdisciplinarity, adequate resources for field
work, scientific rewards geared to practical
achievements (not merely to publications) and
training in the necessary techniques for learn-
ing from farmers.
These models have different strengths and
weaknesses. The farmer-first-and-last model, for
example, is stronger on rhetoric than on the
practicalities of how on-farm research might
link with on-station or laboratory-based
research (indeed, the model scarcely admits
these as legitimate). However, the importance of
both lies in the early impetus they gave to par-
ticipatory research way ahead of the develop-
ment of now widely used participatory rural
appraisal (PRA) techniques.
In organizational terms, the 1970s and
1980s saw a search for ways to re-focus sci-
ence on field problems and opportunities. This
was not to imply that there was no useful role
for reductionist research, simply that the dan-
ger of having station-based and laboratory
research driven by priorities of little interest to
farmers had to be countered. Some observers,
particularly Chambers and Ghildyal, and
Chambers and Jiggins13, were extreme in their
criticism of what they saw as research which
had been outmoded by a growing understand-
ing of farmers' own capacity for experimenta-
tion and by new, rapid methods of diagnosis.
In fact the new methods and perceptions are
not a substitute for reductionist methods, but
are complementary in the insights and focus
they provide, as indicated by the subtitle of the
1987 workshop from which 'Farmer First'14
originated Farmers and Agricultural
Research: Complementary Methods.

3.3.4 Institutionalization of new approaches
Apart from the introduction of farm management
units, the principal means of this re-focusing was
through the creation of multidisciplinary diag-
nostic teams15. There were also efforts to create
institutional links in several dimensions: between
field-oriented diagnostic teams and commodity
or discipline-oriented researchers16, between
research and extension17 and between national
research organizations and the International
Agricultural Research Centres (IARCs).
Writing in 1988, Collinson18 reflected on
the slow pace of institutionalization of FSR in
East Africa, citing various causal factors. He
included vacillations in government and donor
policies, changes in key senior personnel, pater-
nalistic, top-down attitudes and organizational
ethos, inappropriate role models, outmoded
curricula for the training of scientists and poor
reward systems. Even where, as in Zambia, an
institutional mechanism had been created to
inject systems perspectives into commodity
research, the necessary changes in attitudes
and perspective among specialist researchers
were found to take much longer than the simple
mobilization of diagnostic teams.
There can be no doubt that if the numerous
shortcomings identified by Collinson were set
right, the prospects for introducing systems per-
spectives in public sector research would be
improved. But would exclusively public sector
models, such as those in Ethiopia, Kenya and
Zambia, be institutionally sustainable even if all
these conditions were met? The dwindling core
funds for public sector research institutes today
makes them increasingly dependent on donors.
Further, if staff and operating budgets do have
to be cut, the greatest threat is to new initiatives
such as these, and not to the more established
commodity or discipline-based research.
Demands on agricultural research have been
growing over the last decade, at the same time a
core of public research budgets have declined.
These are not simply demands generated by
farmers and those, such as NGOs, who work
with them in the context of a strengthened civil
society. They also include the demand for 'sus-
tainability' of natural resource management,
now invoked at every turn. Within cultivated
land this generates new demands for systems-
based understanding of such externalities as
salinization and chemical pollution, and of how

FSR Understanding Farming Systems 63

interventions relying on joint action, such as
integrated pest management, can best be
designed. However, demands for sustainability
embrace not merely cultivated land but,
increasingly, the management of common
resources such as water, trees and pasture on
land adjoining cultivated areas again, often
demanding group approaches. Agricultural
researchers have long been aware of the need to
draw on such resources in order to ensure, for
example, adequate supplies of fodder and so
contribute to the sustainability of agriculture in
such areas as the mid-hills of Nepal. But they
now have the added burden of devising man-
agement practices and technologies to ensure
sustainable exploitation of this wider natural
resource base. Inevitably, as demands of this
kind broaden, the complexities of diagnosis
increase, giving added impetus to the search for
cost-effective methods.
This growing disparity between expectations
and resources has created an environment
receptive to innovations of two broad, related

* The growing confidence in farmers' own
capacity to identify their needs, has gener-
ated an array of 'farmer participatory
research' (FPR) methods19, the diagnostic
components of which have recently devel-
oped in ways which are often both rapid and
* It is becoming clear to governments that
they no longer have the resources to meet
the wide range of potential research
demands from complex, diverse and risk-
prone areas. There is therefore much talk,
and the beginnings of action, on the cre-
ation of partnerships between government
and private sector organizations, both com-
mercial and non-profit.

3.3.5 FPR and PRA
As the early review by Farrington and Martin in
198821 notes, FPR has its intellectual origins in
the traditions of action research. In practical
terms, participatory methods were first applied
to technical change in agriculture by NGOs and
to special projects to support farmers in identi-
fying the opportunities and constraints they
faced in agricultural development, in meeting

these needs themselves if possible, and with
help from government services if not. FPR was
initially conceived and applied by NGOs in this
'empowering' mode and many early examples
drew on such conventional anthropological
techniques as participant observation. Few,
however, could be classed as 'rapid'.
At the same time, a parallel mode of enquiry
was developing in the form of rapid rural
appraisal22, drawing on efforts to diagnose
farmers' needs without conventional question-
naire surveys or frequent recording techniques,
avoiding their costs and rigidities and the risk of
results so old they miss a 'moving target'23. To
reduce any tendency to be purely extractive,
this model evolved into PRA, the origins of
which were reviewed by Chambers in 199424.
Two general observations need to be made,
and the first concerns the relationship between
PRA and FPR. The large number of case studies
generated by the two has led some to equate
PRA with FPR and Extension (FPR-E). There is,
however, a basic distinction. FPR-E is an
approach to the development of technologies,
embracing diagnosis, screening, testing and
verification that meet farmers' needs. As such,
it is equivalent to FSR-E, but utilizes a wider
range of methods and relies on a wider range of
institutional linkages. PRA is one set of meth-
ods, and has been used primarily as a diagnostic
tool. While it undoubtedly has potential at the
evaluation stage of the research cycle, this
remains largely unexploited. Importantly, it has
little to offer at the experimentation stage.
Efforts to ensure stronger farmer control over
the experimentation process have, for instance,
led to innovations such as the participatory
varietal selection (PVS) highlighted by
Witcombe and Joshi in 199525. Techniques
such as PVS are based on semi-structured inter-
action with farmers over one or more seasons in
which their views on the design and manage-
ment of trials, and their criteria for assessment
of the results, are elicited. Much of the same
kind of interaction had long been used in the
variety of roles played by researchers and farm-
ers in the joint conduct of on-farm research
(OFR). A common theme of the numerous
manuals on OFR, many of which predate PRA,
has been the search for ways of increasing
farmers' control over experimentation. A
wide range of techniques for farmer-to-farmer

64 Chapter 3

extension are being discussed which, again, are
distinct from PRA. These examples demonstrate
that, while PRA has an important contribution
to make to the understanding of farmers' needs,
the concept and practice of FPR-E within a sys-
tems context goes beyond the group of methods
embraced by PRA.
The second observation concerns the differ-
ing objectives, ethos and capacities of NGOs and
government research and extension services. As
noted, early approaches to participation were
introduced by NGOs as part of a broad aim to
empower rural communities. By contrast, gov-
ernment research and extension services view
participation in a functional context as a
means of enhancing efficiency in the design
and uptake of new technologies. There are, of
course, overlaps and intermediate positions: the
better-resourced government services may
engage in more empowering approaches, some
'technology-focused' NGOs may be concerned
more with functional than empowering partici-
pation, and increases in income generated by
functional participation may lead to empower-
ment. But the broad differences in philosophy
and mandate remain and set important limits
on the types and depth of participatory
approach that might reasonably be expected in
government services.

3.3.6 Issues and prospects
For this reviewer, there are four issues in diag-
nostic methods that will occupy centre stage in
the next decade:
* The role of PRA methods.
* The changing role of researchers as FPR
gains ground.
* Multi-agency approaches.
* Expansion pathways.

The role of PRA
PRA has powerfully demonstrated the ability of
village households to contribute to rural devel-
opment planning. It has also generated a sense
of community ownership of development pro-
jects and processes, and a recognition among
administrators that farmer participation
enhances the prospects of success. However, it
is increasingly being seen as a 'new orthodoxy',
and, like all orthodoxy, it is attracting chal-
lenges of diverse kinds. One has to do with intel-

lectual property, viz. the argument that a num-
ber of the methods it embraces predate the term
'PRA. Another is that enthusiasm for methods
has led many to ignore differences in objectives
and in the comparative advantage of different
kinds of organization. It should hardly be sur-
prising that departments of agriculture are
unwilling to become involved in the more
empowering forms of participation, but much
criticism for failing to do so is implicit in the
calls for 'new professionalism'26. There is
increasing concern that enthusiasm for photo-
genic perhaps gimmicky techniques is not
being matched by adequate care in the basics of
unbiased sampling and questioning skills.
Mosse, writing in 199527, sees the need to
develop understanding at several levels of the
organizational and political contexts in which
PRA is conducted. At every level', he says,
'knowledge building, need definition, prioritisa-
tion ... are shaped by social relations, not just
within rural society, but within project teams,
the organizational interests which they are con-
strained to serve and the political environments
in which they work'.
There is certainly scope for further refining
PRA techniques in monitoring and evaluation
and, at the diagnosis stage, in combining them
with techniques for consensus building28. But a
growing concern is the understanding of roles
and process, and, as Alsop et al. argued in
199629. user-friendly techniques have yet to be
developed for these purposes.

The changing role of researchers
The greater involvement of farmers in both the
diagnosis of needs and in decisions over which
technical options to test raises questions over the
future role of researchers. Is it sufficient, as
some would argue, for them merely to act as
facilitators in processes led by farmers and, by
implication, to observe with satisfaction the suc-
cessful adoption of new technologies? Or do they
retain some wider role? Many would argue that
the mandate of researchers remains national (in
some cases, provincial) and that an important
role for them in FPR should be to identify not
simply what works, but why and how so that
'baskets of choices' potentially relevant to other
areas can be assembled. Thus. there is both con-
tinuity and change in the role of researchers:
they continue to be concerned with identifying

FSR Understanding Farming Systems

what is potentially relevant to wider 'research
domains', but now assemble information to feed
into these more from on-farm observations than
from on-station experiments.

Multi-agency approaches
It is argued above that governments will tend to
adopt functional approaches to participation, in
contrast with the empowering approaches of
NGOs. Nowhere is the contrast starker than in
the different approaches to technologies requir-
ing 'joint action' by the two types of organiza-
tion. Some farm activities, including various
IPM techniques, require joint action if they are
to be effective and the management of common
pool resources can only be performed by farmer
groups. However, the record of government
staff in forming anything more than very tem-
porary groups is poor30.
NGO approaches are much smaller in scale,
more time-consuming and more intensive, rely-
ing on the development of a sense of commu-
nity identity, and of leadership, participation
and conflict resolution skills. The groups formed
with NGO support are intended to become self-
sustaining, to address their own community
needs where possible and to make demands on
government services where necessary. These
objectives are not always achieved, of course,
but the needs in agriculture and natural
resource management are identified and
addressed in this context. A particular problem
facing NGOs is their limited awareness of the
range of technical options available to meet
farmers' needs, and limited access to such
options. Proposals now abound for multi-
agency approaches seeking to combine the
strengths of NGOs in needs assessment and
group formation with those of government in
developing technical solutions.
The views and mandates of government and
NGOs diverge to some degree, and the area of
shared aspirations is limited. This does not imply,
however, that NGO ideologies will necessarily be
'tainted' if they work with government. A very
wide range of interaction has been identified31
and even the most empowerment-oriented NGOs
need not find it difficult to make technical
demands on government. A widespread problem
is that government services have historically
been driven more by an ethos of delivering pro-
grammes than of responding to demands.

Techniques based on process documentation and
process monitoring are now being developed to
assess how far joint activities proceed in line with
expectations, and to bring them back on track
where necessary32. The further development of
multi-agency approaches and methods of moni-
toring them will clearly be a key influence on the
type and extent of future multi-agency
approaches, including the articulation by NGOs
of farmers' needs to government services.

Expansion pathways
At the risk of caricature, NGO approaches
towards diagnostic methods may be character-
ized as 'deep but small-scale', and those of gov-
ernment services as 'large-scale but superficial'.
The intensive, face-to-face nature of NGO
approaches has made it difficult to spread the
approaches, but the strengths of both types of
institution might be combined. Multi-agency
approaches are not the sole solution. Access to
powerful new media (television, video, local
radio) is expanding rapidly in many rural areas.
With a little imagination, improved use of these
could lead to advances in diagnosis in two ways:
* The capacity of these media to enhance
farmers' awareness of technologies that
have worked well in areas similar to their
own. This may stimulate 'self diagnosis', vis-
its to other areas and testing and adoption at
minimal cost to government or NGO.
* Their capacity to convey, in farmers' own
words, how they set about forming groups
and pressed their demands on NGOs or gov-
ernment services could have important
demonstration effects.
As noted, changes are needed on both sides
before such approaches will become widespread.
Many regard direct interaction between
researchers and farmers' associations as the
most sustainable arrangement for the long term.
However, only in commercial crops have associ-
ations (usually commodity associations) proven
strong enough to influence research agenda33
and there are grave doubts over whether small
farmer groups can be created and sustained in
semi-subsistence contexts without long-term
external support, possibly by NGOs34. Moreover,
even in the north, the extent to which farmers'
organizations have been willing to finance
research in the face of the withdrawal of public

Chapter 3

funding has been limited35. Further options for 0 In others, NGOs may support local member-
change include the following: ship organizations in identifying or express-

* Where public sector research and extension ing needs, or may do so on their behalf.
services are strong, they can be expected to Overall, what is clear is that there will be no sin-
introduce participatory appraisal on a wide gle pattern for the introduction and wide imple-
scale, providing that appropriate reward mentation of improved diagnostic methods in
structures and funding mechanisms are the future. Local agroecological and socioeco-
introduced. nomic settings will increasingly determine the
* In some settings, farmers' associations may choice of methods, and institutional configura-
be able to articulate their members' require- tions will have to be tailored to local settings.
ments to research organizations.


1. Byerlee, D. & A. Siddiq, 1994. Has the Green Revolution been sustained? The quantitative impact of the
seed-fertiliser revolution in Pakistan revisited. World Development. 22(9). 1345-61.
2. Norman, D.W., 1974. Rationalising mixed cropping under indigenous conditions: the example of north-
ern Nigeria. Journal of Development Studies, 11, 3-21.
3. Collinson, M.P., 1972. Farm Management in Peasant Agriculture. Praeger, New York.
4. Ruthenberg, H., Ed., 1968. Smallholder Farming and Smallholder Development in Tanzania. Weltforum
Verlag, Munich.
5. Mellor, J.W., 1966. The Economics of Agricultural Development. Cornell University Press. Ithaca. NY.
6. Gilbert, E.H., D.W. Norman & EE. Winch, 1980. Farming Systems Research: A Critical Appraisal. MSU
Rural Development Paper no. 6. Michigan State University, East Lansing.
7. ISRA, 1977. Recherche et D6veloppement Agricole: les Unit6s Experimentales du Senegal. Institute
S6n6galais du Recherche Agronomique, Dakar.
8. Hildebrand, P.E., 1981. Combining disciplines in rapid rural appraisal: the 'Sondeo' approach.
Agricultural Administration, 8(6), 423-32.
9. Collinson, M.P.,1972. (Op. cit.).
10. IDS, 1979. Rapid Rural Appraisal. A conference held at the Institute of Development Studies, December.
Mimeo Proceedings. Institute of Development Studies, Brighton, UK.
11. Rhoades, R.E. & R.H. Booth, 1982. Farmer-back-to-farmer: a model for generating acceptable agricul-
tural technology. Agricultural Administration, 1, 127-37.
12. Chambers, R. & B.P. Ghildyal, 1985. Agricultural research for resource-poor farmers: the farmer-first-
and-last model. Agricultural Administration. 20, 1-30.
13. Chambers, R. & J. Jiggins, 1986. Agricultural research for resource-poor farmers: a parsimonious para-
digm. IDS Discussion Paper 220. IDS, University of Sussex.
14. Chambers, R., A. Pacey & L.A. Thrupp, 1989. Farmer First: Farmer Innovation and Agricultural
Research. IT Publications, London.
15. Gait, D.L. & S.B. Mathema, 1987. Farmer Participation in Farming Systems Research. FSSP Networking
Paper no. 15.
16. Kean, S. & L.P. Singogo, 1968. Zambia: Organisation and Management of the Adaptive Research
Planning Team (ARPT), Research Branch, Ministry of Agriculture and Water Development. OFCOR Case
Study no. 1. ISNAR, The Hague.
17. ICAR, 1988. Report of the ICAR Review Committee. Indian Council for Agricultural Research. New Delhi.
18. Collinson, M.P., 1988. The development of African farming systems: some personal views. Agricultural
Administration and Extension, 29(1), 7-22.
19. Farrington, J. & A.M. Martin, 1988. Farmer Participation in Agricultural Research: a review of con-
cepts and practices. Agricultural Administration Unit Occasional Paper no. 9. Overseas Development
Institute, London.
Okali, C., J. Sumberg & J. Farrington, 1994. Farmer Participatory Research: Rhetoric and Reality. IT
Publications. London.
20. Chambers, R., et al., 1989. (Op. cit.).
Scoones, I. & J. Thompson, 1994. Beyond Farmer First. IT Publications, London.
21. Farrington, J. & A.M. Martin, 1988. (Op. cit.).

FSR Understanding Farming Systems

22. IDS, 1979. (Op. cit.).
23. Maxwell, S., 1986. Farming systems research: hitting a moving target. World Development, 14(1),
24. Chambers, R., 1994. The origins and practice of participatory rural appraisal. World Development,
22(7), 953-69.
25. Witcombe, J. & A. Joshi, 1995. Farmer Participatory Approaches for Varietal Breeding and Selection,
and Linkages to the Formal Seed Sector. Paper presented at the IDRC Workshop on Participatory Plant
Breeding at the International Agriculture Centre, Wageningen, 24-30 July, 1995.
26. Pretty, J.N. & R. Chambers, 1993. Towards a Learning Paradigm: New Professionalism and Institutions
for Agriculture. IDS Discussion Paper no. 334. Institute of Development Studies, Brighton.
27. Mosse, D., 1995. 'People's Knowledge' in Project Planning: The Limits and Social Conditions of
Participation in Planning Agricultural Development. Agricultural Research and Extension Network Paper
no. 58. Overseas Development Institute, London.
28. Warner, M., 1997. 'Consensus' Participation: an Example for Protected Areas Planning. Public
Administration and Development. 17(4), 413-32.
29. Alsop, R.G., R. Khandelwal, E.H. Gilbert & J. Farrington, 1996. The human capital dimension of col-
laboration among government, NGOs and farm families: comparative advantage, complications and obser-
vations from an Indian case. Agriculture and Human Values, 13(12), 3-12.
30. Chowdhury, M.K. & E.H. Gilbert, 1996. Reforming Agricultural Extension in Bangladesh: Blending
Greater Participation and Sustainability with Institutional Strengthening. Agricultural Research and
Extension Network Paper no. 61. Overseas Development Institute, London.
31. Farrington, J. & A.M. Martin, 1988. (Op. cit.).
32. Alsop, R.G. et al., 1996. (Op. cit.).
33. Bebbington, A., D. Merrill-Sands & J. Farrington, 1994. Farmer and Community Organisations in
Agricultural Research and Extension: Functions, Impacts and Questions. Agricultural Research and
Extension Network Paper no. 47. Overseas Development Institute, London.
Carney, D., 1996. Formal Farmers' Organisations in the Agricultural Technology System: Current
Roles and Future Challenges. Natural Resources Perspectives, no. 14. Overseas Development Institute,
34. Mosse, D., 1996. Local Institutions and Farming Systems Development: Thoughts from a Project in Tribal
Western India. Agricultural Research and Extension Network Paper no. 64. ODI, London.
35. Pollock, C., 1995. Reconciling the Irreconcilable? The Effective Management of Agricultural Research.
Paper presented to the ICAR/ODA Workshop on Research for Rainfed Farming, September 1995 at Central
Research Institute for Dryland Agriculture, Hyderabad, India.

Hilary Sims Feldstein

Development specialists have realized that ignoring women in the development process has meant ignor-
ing a positive force for change. Increasing women's productivity has substantial payoffs for family liveli-
hoods, including the better health and education of the next generation.

3.4.1 Introduction Initially, researchers saw technology as 'neu-
tral'. They came to understand that acceptance
The impact of FSR on agricultural research has could be size biased and that the constraints of
been twofold. It showed that:
low resource farmers differed from those of
* Agricultural production was one part of a larger landholders. FSR focused attention on
complex system guiding farmers' choices. farmers and their diverse circumstances.
* Farmers had a sophisticated understanding However, a blind spot remained. While technol-
about their current system and the costs and ogy was seen as size biased, it was not seen as
benefits of new technology, including specific gendered. Too frequently, learning about and
preferences which governed its acceptability, from farmers meant learning about and from

68 Chapter 3

male farmers. Women farmers and their knowl-
edge and preferences about agricultural pro-
duction were virtually invisible. Early guidelines
to FSR did not include the analysis of gender-
disaggregated data as a source of useful insights
for scientists. Most FSR practitioners considered
the household a 'black box' in which resources,
responsibilities and benefits are equally (or opti-
mally) shared. They assumed that one individ-
ual, usually male, could speak for all.

3.4.2 The importance of gender

In the 1930s and 1940s, a number of anthro-
pologists, trained in careful observation of the
daily lives of those they studied, detailed the
productive activities of women and men1. But
this kind of social analysis did not, in general,
carry forward into FSR. Boserup's 1970 study,
'Women's Role in Economic Development'2,
reopened the case. She stressed the important
contribution made by women to agricultural
production, drawing attention to their different
role according to population density, the inten-
sity of production and the availability of hired
labour. Modernization and intensification of
agriculture would mean increased labour by
both men and women. She pointed out that
shared labour did not necessarily mean shared
benefits; men were more likely to predominate
with their access to markets and new technolo-
gies. Indeed, women's income and position
could be at risk.
Beginning in 1975, the launch of the
Decade for Women, researchers associated with
US land grant universities working in develop-
ing countries provided abundant examples of
women's roles in agriculture, including Horn
and Nkambule-Kanyima in 19843. Women
often had exclusive responsibility for particular
crops, livestock or operations. Their roles and
preferences were sometimes complementary,
sometimes parallel and sometimes in conflict
with those of men4. Case studies from
Indonesia showed how improved technologies
could actually harm women's opportunities and
welfare. The introduction of rice mills operated
generally by men, replaced the hand pounding
of rice carried out by women. The result was
the loss of employment and income by poor
women labourers.

In an early study of irrigated rice produc-
tion in Gambia, Jennie Dey described how the
assignment of irrigated plots to men trans-
ferred the ownership of rice production and its
income from women to men5. In 1986,
Christine Jones's widely circulated study in
Cameroon showed how failing to take women's
interests into account led to serious miscalcula-
tions in expected productivity and economic
returns6. Both demonstrated that efficiency
and equity were at risk when research or devel-
opment failed to take account of women's roles
in agriculture. There was increasing evidence
that households were complex decision-making
entities7 and this sparked several efforts to
combine improved understanding of house-
hold dynamics with FSR.

3.4.3 Strategies in the introduction of gender
analysis into FSR

The inception and development of attention to
gender issues within FSR are like a braid, made
up of several strands that overlap and
strengthen each other. These strategies were
frequently opportunistic, building on existing
openings and developing momentum. They fall
largely into four categories:
* Conferences bringing together those study-
ing intra-household decision making (princi-
pally anthropologists), and social scientists
engaged in FSR (mainly economists)8.
* The development of methods for use by FSR-
E researchers to gain a better understanding
of gender roles, and incorporation of these
methods into the conduct of FSR-E9.
* Specific studies to show how FSR could
incorporate an intra-household perspective
with useful results.
* Conferences and workshops that gave visibil-
ity to individual projects where knowledge of
women's role had led to improved project
These legitimized gender analysis and became
the vehicles for the discussion and develop-
ment of reliable gender-sensitive research
methods. Three of the main training pro-
grammes for FSR were: CIMMYT's East
African Economics Program based in Nairobi,
Farming System Support Project's (FSSP)
assistance to FSR projects funded by USAID.

FSR Understanding Farming Systems

and the IRRI Farming Systems Training
Program. Each of these came to accept the
legitimacy of gender analysis and incorpo-
rated it into its training programmes. Looking
back, the perseverance of Kate McKee of the
Ford Foundation in identifying and funding
different opportunities for furthering this issue
was crucial. The Ford Foundation provided
long-term support to the Intra-Household and
FSRE Case Studies Project (IHH/FSR-E) from
1984 to 1994 and the Women in Rice
Farming Systems (WIRFS) network from the
mid 1980s to the mid 1990s the two most
persistent efforts to develop methods and
engage FSR researchers.

The award of a large technical assistance and
training contract to the FSSP to the University
of Florida (UF)
This provided the means for stronger links
between US land grant universities and FSR-E.
Research by the Population Council of New
York City showed that where women had a
stronger bargaining position, fertility appeared
to drop and women's own productive activities
significantly enhanced their positions. Judith
Bruce, with the support of the Ford Foundation,
began a series of studies to help development
planners understand women as producers.
Constantina Safilios-Rothschild conducted stud-
ies comparing women's involvement in agricul-
tural production with their invisibility in official
figures. And a group from Harvard constructed
a conceptual framework for a set of case studies
for the World Bank and USAID, now known as
the Harvard framework"1.

WIRFS Conference at IRRI
The WIRFS Conference at IRRI, supported by
the Ford Foundation, was an early effort to
explore the links between women and FSR. The
meeting gave birth to the IHH/FSR-E Case
Studies Project. Shortly after the conference,
IRRI created the WIRFS Network, also funded
by Ford and led by Gelia Castillo, Professor at
the University of the Philippines at Los Bafios,
and Virgilio Carangal, head of the Asian Rice
Farming Systems Network (ARFSN). The
Women in Agricultural Development (WIAD)
programme began to meet on an informal basis
at the UF12.

3.4.4 Milestones and achievements

Joyce Moock at the Rockefeller Foundation and
Kate McKee from Ford organized the Bellagio
Conference on Understanding Africa's Rural
Households and Farming Systems to address
gender issues with the leading farming systems
practitioners'1. This conference provided the
impetus to find practical tools for incorporating
intra-household and inter-household analysis
into FSR. The IHH/FSR-E Case Studies Project
was initiated, led by Hilary Sims Feldstein and
Susan Poats and, again, funded by the Ford
Foundation. The project developed seven train-
ing case studies integrating intra-household
considerations and FSR14 and a handbook on
gender-sensitive research methods1 5. The pro-
ject also organized 'gender methodologies pan-
els' at the annual FSR-E meetings.

The Bellagio Seminar on Women and
Agricultural Technology: Relevance for
Research, sponsored by the Rockefeller
Foundation and the International Service for
National Agricultural Research (ISNAR),
brought together the 12 international centres
of the CGIAR16. The Conference agreed that
gender was an important variable in agricul-
tural research and that centres should recog-
nize this by linking gender directly to
technology generation and use.
The WIRFS Network was formed. At its
first workshop, the priority was to build some
regional experience. The network leader,
Thelma Paris, an associate scientist at IRRI,
used a judicious mixture of conferences, small
research grants, individual mentoring and
training, to encourage scientists from national
programmes to incorporate gender analysis into
their FSR.

The monograph by Janice Jiggins in the CGIAR
study on gender-related impacts and the work
of the IARCs highlighted the accumulating evi-
dence of women's substantial contribution to
agricultural production and to household nutri-
tion17. Jiggins was critical of the Centres' persis-
tent gender bias that contributed to the
invisibility of women in FSR programmes. The

Chapter 3

UF/WIAD Conference on Gender Issues in
Farming Systems Research and Extension took
place with the support of three departments at
UF, and the Ford and Rockefeller Foundations.
Over 350 participants from all regions of the
globe attended'8.

The CIMMYT Networkshop on Household
Issues and FSR was organized by Allen Low,
Alistair Sutherland, Feldstein and Poats in
Lusaka, Zambia, in April, 1987. Zambia was
already a leader among national programmes
in taking steps to incorporate a gender perspec-
tive in its research process. Biological and social
scientists from nine national programmes spent
5 days discussing the IHH/FSR-E Zambia case,
presenting papers and visiting the field19. This
laid the groundwork for the incorporation of
gender concerns into CIMMYT's East and
Southern Africa FSR training and a wider
appreciation of these concerns in general.

The WIRFS conference, IRRI brought together
national researchers from seven south-east
Asian countries. Some, with small grants from
the 1985 meeting, had applied a gender per-
spective to their ongoing farming systems work.
For others, this workshop was their first expo-
sure to gender analysis, widening the reach of
the WIRFS programme in south-east Asia. In
September, Feldstein and Flora provided a day of
training on gender and FSR-E for the
International Development Research Center
(IDRC), reaching another major set of FSR
actors. Tangible evidence that the argument for
gender analysis had won recognition as good
social analysis came when the very FSR practi-
tioners who had baulked at using the words
'women' and 'gender' at the inception of the
Case Study Project, suggested that gender issues
be a subtheme of the 1988 conference.

IDRC published the monograph 'The Gender
Variable in Agricultural Research' by Feldstein,
Flora and Poats in English, French and Spanish
and gave it wide free distribution20. Later in the
year 'Working Together: Gender Analysis in
Agriculture', by Feldstein and Poats, was pub-
lished21. Gender became a subtheme at the

annual FSR-E conference, and sessions on
intra-household analysis became the best

Attending the WIRFS Conference in Bogor,
Indonesia, were senior agricultural administra-
tors from India, Indonesia, Nepal, the
Philippines and Thailand. Using the GAP con-
ceptual framework, the researchers who had
received small grants and training from WIRFS
in 1988 presented their analyses of the sites in
which they were working. The administrators'
reaction was very positive and they endorsed a
continuation of the WIRFS project and a
request that IRRI do more to train their
national scientists in gender analysis. IRRI sub-
sequently incorporated gender analysis into its
farming systems training as well as providing a
separate course on gender analysis. Through
the sustained commitment of IRRI and the Ford
Foundation, the leading donor, the programme
flourished for 10 years. During this period
WIRFS worked with 23 organizations from
National Agricultural Research Systems
(NARS) in nine south and south-east Asian
countries. WIRFS also began to have an impact
on technical research conducted at IRRI22.
By 1990 FSR was losing favour with donors
but gender as a development variable was
receiving more attention. In 1991, CGIAR
launched its Gender Program. Beyond this,
Agenda 21 explicitly discussed women's
reliance on the environment for their liveli-
hoods and the importance of involving women
in sustainable natural resource management
(NRM). The focus on NRM has provided a new
venue for applying the principles of FSR-E and
with it gender analysis, but also raises new

3.4.5 Evolution of methods for collecting
gender-sensitive and disaggregated data
The rubric, gender analysis, is often used to
refer to both the conceptual framework, used
for analysis and planning, and a range of
methods for the collection of gender disaggre-
gated data from both men and women farmers.
The conceptual framework for gender analysis
in FSR-E is an aid to understanding the pattern
of roles and responsibilities in the farming sys-

FSR Understanding Farming Systems

tem. The framework focuses the gathering of
data and its analysis on four key areas: activi-
ties, resources, benefits and inclusion. The cen-
tral question is 'Who does what?' Data is
gender disaggregated in its collection and its
analysis and information is gathered, arranged
and compared on the activities of women and
men farmers.
One of the major contributions of the frame-
work is the use of a gender-disaggregated farm-
ing systems calendar. This shows when the
labour constraint is most severe, and what and
whose tasks would be affected by any changes
in the farming system23. In Burkina Faso, for
example, a linear programming of sex-specific
labour inputs explained the limited extent of
adoption of tied ridges, a technology for mois-
ture conservation. Women's labour was con-
strained by their household responsibilities or
involvement in 'own account' enterprises such
as beer making or working in their own fields24.
To predict the availability of other resources,
researchers use the framework to analyse data
on men's and women's access to, and control of,
resources and the benefits from production. By
comparing the costs and benefits of technolo-
gies by gender, researchers can anticipate the
likely constraints to uptake. An analysis of ben-
efits also includes an understanding of the
shared and differing preferences of men and
women for specific traits in a crop. Finally, mon-
itoring the inclusion of women and men in the
different stages of FSR enhances researchers'
awareness of points at which gender-specific
knowledge can usefully be brought to bear.
While the focus of the framework is on gender
and age, other important variables ethnicity,
class, life cycle stage could also be incorpo-
rated into the analysis.
When the IHH/FSR-E Case Studies Project
began, it followed the current methodology of
FSR-E diagnosis through informal and formal
surveys, planning and design, on-farm experi-
mentation, evaluation and dissemination. Each
of these stages needed an expansion of method-
ology to ensure the collection of data about
women as well as men. Attention to gender in
informal diagnostic surveys means widening
both the kinds of informants (more women)
and the key questions. In Botswana and
Zambia, researchers included female heads of
household among those surveyed. In Sta

Barbara in the Philippines, women members of
joint households were interviewed separately in
a resurvey that provided a better description of
the farming system and respective roles. By
interviewing women not included in the origi-
nal surveys, researchers discovered that the
processing of glutinous rice was an important
source of household income. This predomi-
nantly women's activity became a fertile area
for experimentation with improved technolo-
gies, from new varieties grown by men, to new
post-harvest machinery used by women25. In
Peru, work in the community revealed, unex-
pectedly, that women, rather than men, were
responsible for the care of animals. Through
discussion with the women, scientists learned
that they did not use chemical dipping because
of its high cost, and that there were local treat-
ments for parasites. Scientists helped to orga-
nize a production research group to conduct
trials based on local plants and chemical dip-
ping. A local leaf proved to be as effective as the
chemical, but it was scarce. This prompted the
women to begin plant multiplication to over-
come this limitation26.
On-farm trials often required changes better
to capture the practices and assessments of
both women and men27. In Botswana and
Zambia, female heads of households were also
collaborators in on-farm trials and researchers
held separate field days to ensure they heard the
candid opinions of women farmers. In
Zambia28, researchers collected and analysed
gender disaggregated data which proved partic-
ularly useful in technology surveys, as
Sutherland wrote in 1994: 'For example, in an
assessment of intercropping in Lusaka
Province, male and female respondents saw dif-
ferent advantages and disadvantages. ... Certain
treatments were ranked high only by the
women respondents. This helped in the design
and targeting of subsequent on-farm experi-
Including women as sources of information
and partners in FSR continues to present
methodological challenges. In many societies,
men dominate the 'public domain'; women may
not be present and when they are present they
do not speak. Local custom may prevent men
from questioning them individually. Group
interviews and the use of more women field
researchers have helped to overcome this

Chapter 3

constraint. Researchers may have to shell beans
with women in the kitchen30 or engage with
separate focus groups31 to hear women's voices.
Understanding the general pattern of the roles
of women and men may require very specific
questions about the details of particular pro-
duction activities.
In the late 1980s, FSR researchers and oth-
ers began to expand the tool kit for learning
from farmers to include mapping, focus groups,
transepts and matrix ranking. Though still
uneven in practice, these techniques prompted
a shift away from extracting data to design solu-
tions back at the research station towards more
fully incorporating farmers in decision-making
about what should be tried and how to test it. To
use these tools, scientists concerned with gen-
der and other variables have made adaptations
to ensure that the voices of women and men are
heard. This is especially true in the work of the
Ecology, Community Organization, and Gender
(ECOGEN) project at Clark University, the
Sustainable Agriculture and Natural Resource
and Environment Management Collaborative
Research Support Project (SANREM CRSP), and
UF's project on Managing Ecologies and
Resources with a Gender Emphasis (MERGE).
As the participatory tools proliferate, adap-
tation to include gender analysis has two impor-
tant elements. The first, when undertaking
community mapping or transepts, is to ask
questions about each element of the landscape,
its different uses, ownership, and who has
responsibility for each enterprise32. Activities,
resources and benefit analyses can be con-
ducted using cards with drawings of men,
women and children and different enterprises.
Using such cards, a group (all men, all women,
or mixed) can collectively sort out men's and
women's roles and responsibilities for each
enterprise33. Participants can explain benefits
by breaking down plants by their many prod-
ucts and the ownership, responsibility and ben-
eficiary of each34.
Second, to ensure that women's opinions are
registered, researchers should conduct inter-
views with sensitivity. They must be careful
about how questions are worded and what
questions are asked, focusing on the areas
where they know the women to be expert, for
example, or conduct separate interviews keep-
ing in mind that all women are not the same.

These same principles may apply to groups dif-
ferentiated by other variables such as class or
ethnicity or age. 'Questions of Difference', a
training video produced by Irene Guijt of the
International Institute for Environment and
Development (IIED) shows adaptations of par-
ticipatory methods and the value of informa-
tion from women to NRM3'.

3.4.6 Evidence on making a difference:
efficiency, welfare, equity and empowerment
A question often asked is 'Does it make a differ-
ence?' Does understanding the gender dimen-
sion in division of labour, access and control of
resources, responsibility and decision-making
make a difference? The application of gender
analysis to agricultural research is new, but
there are now many examples of where it has
made a difference. The answer to this question
depends on the purpose to which the analysis is
put. Because women's and men's different roles
and interests may weigh heavily in the adoption
of improved technologies, FSR leaders now rec-
ognize that gender differentiation is 'good' FSR.

One reason for its use is efficiency. In Rwanda,
for example, scientists learned from on-farm tri-
als that it was the women who were the pre-
dominant bean growers. In an experiment on
farmer evaluation, women bean experts came
on-station to discuss 21 varieties in the final
stages of selection with breeders. This was
much earlier in the breeding cycles than farmer
evaluation of the few varieties moved forward
into on-farm trials. The women experts selected
for very specific traits; for planting where there
is wind. for planting near bananas, and three or
four varieties to plant in their own on-farm
experiments. Their selections outperformed
their own check varieties more frequently and
by a greater margin than had breeders' selec-
tions in similar trials36.
Gender analysis is used for an ex ante assess-
ment of the fit of the technology into the farm-
ing system. It helps identify the appropriate
collaborators for specific operations or enter-
prises. Gender analysis also helps identify who
has local and practical knowledge about differ-
ent parts of the farming system and of the nat-
ural resources upon which it draws. As in the

FSR Understanding Farming Systems 73

Sta Barbara case, this approach can lead to
identifying overlooked areas where researchers'
knowledge and experimentation can make a
real contribution.

Welfare and equity
Gender analysis makes a difference in address-
ing issues of welfare and equity. It can be used
ex ante to anticipate whether women's condi-
tions are worsened or improved by a new tech-
nology. Concerns about displacing female hired
labour have featured in the research by the
International Crops Research Institute for the
Semi-Arid Tropics (ICRISAT) on herbicides and
have influenced the decision by the International
Center for Agricultural Research in the Dry
Areas (ICARDA) in developing a lentil
harvester37. In Zambia, women's concerns
about losing control of their crop negated a pro-
posed intercropping of women's beans with
men's maize38. Used more proactively, gender
analysis provides a means for identifying those
specific activities by women where research or
extension can produce tangible benefits. For
subsistence farmers, research to reduce weed
populations would reduce women's drudgery
and IRRI has engineered several machines
scaled for women and women's tasks in produc-
tion and post-harvest processing. In both cases,
what is drudgery reducing in the case of subsis-
tence farming may be labour displacing on
larger farms. Such situations require sensitive
analysis and exploration of options. The equity
orientation is currently receiving increased
attention. Development specialists have realized
that ignoring women in the development
process has meant ignoring a positive force for
change. Increasing women's productivity has
substantial pay-offs for family livelihoods,
including the better health and education of the
next generation39.

A third area is women's empowerment,
strengthening their ability to make choices
about their own livelihoods. Researchers work-
ing with sheep in the Andes, where livestock
were women's responsibility, helped form a
women's sheep association to discuss research
needs and to design trials. This forum became
the vehicle for discussing many other issues and
to formalization of the group as a place to dis-

cuss women's and community affairs40. In
Guimba in the Philippines, an improved rice
mill reduced costs and time for processing rice.
The arrangements by women to run the mill
became the basis for a women's association
which eventually gave them a stronger voice in
community affairs41.
Whatever the intention efficiency, equity
or empowerment gender analysis provides a
'map' for identifying men's and women's roles
and helps improve the choice and design of
improved technologies.

3.4.7 Continuing challenges
The application of FSR to NRM
NRM shifts the emphasis from the farm house-
hold to the community or communities in a
larger landscape. Several elements may limit
attention to gender: more complexity; more and
different kinds of stakeholders, and, at the pub-
lic community or intercommunity level, women
may be invisible. While gender may not be the
dominant variable in differentiating stakehold-
ers, it is still important.

The role of women researchers
Male researchers and administrators often
claim that adding women to the team will 'take
care of the gender question'. A woman
researcher may find it easier to reach rural
women and her own experiences may help her
understand their circumstances. However,
being a woman does not automatically qualify
one to carry out gender analysis. This is a skill
that must be learned, by women as well as men.

The circumstances for women professionals
More professional women come to gender
analysis workshops than to any other kind of
FSR forum. For many, such a workshop is the
first event of its kind that they have attended.
Away from the main proceedings, they discuss
the difficulties of their workplaces. Some are
wholly accepted as members of the team, but
many are ignored, belittled, sexually harassed
and find their chances of career advancement
extremely limited. The workplace rarely takes
account of the greater time constraints on
women than men as women continue to bear the
greater responsibility for household production.
There is a disturbing parallel between women's

74 Chapter 3

treatment in an organization and the serious-
ness with which the organization addresses the
needs of women as clients.

The capacity to implement gender-sensitive
The comparative advantage for gender analysis
in FSR is with national programmes and NGOs,
but with decreasing funds for agricultural
research and development, the capacity for FSR
is also shrinking. Some national programmes
are now making a concerted effort to build gen-
der analysis into their research, but many lack
the resources or the commitment.

Resistance from men
However reasoned the argument for using gen-
der analysis to improve the efficiency of
research, many male scientists are still inclined
to not listen, or to laugh or to ignore. Often the
question 'Will it make a difference to what I am
working on?' reflects their reluctance even to
read the basic texts and consider their applica-
tion to their own work. Their own socialization
on women's roles at home often makes it diffi-
cult for them to take women seriously as col-
leagues or as sources of information. There are
now a number of men who speak out forcefully

for gender analysis and the importance of
addressing women's needs, including the editor
of this book. But in the long run it is practice
and interventions at the field level which make
a difference42. Given entrenched habits and atti-
tudes, changes are slow.

3.4.8 Conclusion
Fifteen years of work in gender analysis has
paid off in the increasing legitimacy and use of
gender analysis in FSR-E and other agricultural
research. Furthermore, gender analysis, by its
focus on different kinds of farmers and their
interactions, has made anthropological insights
more visible in FSR, and has done so in a way
which adds clarity to the presentation. A num-
ber of national programmes and NGOs in Africa
and Asia are giving more attention to women as
farmers and as fellow scientists. Women them-
selves are increasingly speaking out on their
specific needs for technology, and the credit and
extension services that enhance their produc-
tivity. However, those committed to excellence
and equity in FSR will need to be persistent in
ensuring that gender is fully considered in tech-
nology development.

1. Richards, A.I., 1961. (Reprint; original 1939) Land, Labour and Diet in Northern Rhodesia. Oxford
University Press, London.
2. Boserup, E., 1970. Women's Role in Economic Development. St Martin's Press, New York.
3. Horn, N. & B. Nkambule-Kanyima, 1984. Resource Guide: Women in Agriculture, Botswana. The
Bean/Cowpea Collaborative Research Support Program, 1 October 1984. Michigan State University.
4. Cloud, K., 1985. Women's Productivity in Agricultural Systems: Considerations for Project Design. In
Overholt et al., (Eds) Gender Roles in Development Projects. Kumarian Press, West Hartford, Connecticut.
5. Dey, J., 1981. Gambian Women: Unequal Partners in Rice Development Projects? In: Nelson. N. (Ed.)
African Women in the Development Process. Frank Cass, London.
6. Jones, C.W., 1986. Intra-household bargaining in response to the introduction of new crops: a casestudy
from North Cameroon. In: Moock (Op. cit.).
7. Guyer, J.I., 1980. Household budgets and women's incomes. Paper presented at the Symposium on
Women in the Work Force, American Anthropological Association Meetings. 1979.
8. Cloud, K., 1985. (Op. cit.).
9. (Ibid.)
Feldstein, H. Sims & S.V. Poats (Eds), 1989. Working Together: Gender Analysis in Agriculture, Volume
1. Kumarian Press, West Hartford, CT.
Feldstein, H. Sims & J. Jiggins (Eds), 1994. Tools for the Field: Methodologies Handbook for Gender
Analysis in Agriculture. Kumarian Press, West Hartford, CT.
Thomas-Slayter, B., A. Lee Esser & M. Dale Shields, 1993. Tools of Gender Analysis: A Guide to Field
Methods for Bringing Gender into Sustainable Resource Management. Clark University. Worcester. MA.

FSR Understanding Farming Systems

10. Ashby, J., 1989. Production and Consumption Aspects of Technology Testing in Pescador. In H. Sims
Feldstein & S.V. Poats (Eds) Working Together: Gender Analysis in Agriculture, Volume 1. Kumarian Press,
West Hartford, CT.
Paris, T.R., 1989. Philippines: Women in Rice Farming Systems, Crop-Livestock Project, Sta Barbara,
Pangasinan. In: H. Sims Feldstein & S.V. Poats (Eds) Working Together. (Op. cit.).
Shinawatra, B., 1988. The incorporation of women's concerns in the design and impact assessment of a
FSR/E project: A case of Northern Thailand. Paper presented at the 1988 Farming Systems Research and
Extension Symposium, Fayetteville, Arkansas, 9-12 October, 1988.
11. Overholt, C., M.B. Anderson, K. Cloud & I.E. Austin (Eds), 1985. Gender Roles in Development
Projects. Kumarian Press, West Hartford, CT,
12. Poats, S.V., M. Schmink & A. Spring (Eds), 1988. Gender Issues in Farming Systems Research and
Extension. Westview Press, Boulder, CO.
13. Moock, J. Lewinger (Ed.), 1986. Understanding Africa's Rural Households and Farming Systems.
Westview Press, Boulder, CO.
14. Feldstein, H. Sims & S.V. Poats (Eds), 1989. (Op. cit.).
15. Feldstein, H. Sims & J. Jiggins (Eds), 1994. (Op. cit.).
16. Rockefeller Foundation and ISNAR, 1985. Women and Agricultural Technology: Relevance for
Research. Volume 1: Analyses and Conclusions. Volume 2: Experiencies in International and National
Research. Report for the CGIAR Inter-Center Seminar on Women and Agricultural Technology. Bellagio,
Italy, 25-29 March, 1985. The Hague, Netherlands.
17. Jiggins, J., 1986. Gender-related impacts and the work of the international agricultural research centers.
CGIAR Study Paper No. 17. The World Bank, Washington, DC.
18. Poats, S.V., M. Schmink & A. Spring (Eds), 1988. (Op. cit.).
19. Sutherland, A. (Ed.), 1987. Report on a Networkshop on Household Issues and FSR. Lusaka, Zambia.
Networking Workshop Report, No. 10. CIMMYT, Harare.
20. Feldstein, H. Sims & S.V. Poats (Eds), 1989. (Op. cit.).
21. (Ibid.).
22. CGIAR Gender Program, 1994. Partners in Selection: Women Bean Experts and Bean Breeders.
Consultative Group on International Agricultural Research, Washington, DC.
Chater, S. & V. Carangal, 1996. On Farmers' Fields: Portrait of a Network. Manila: IRRI.
Paris, T., 1995. Women in Rice Farming Systems Network: Progress Report, 1991-95 (mimeo).
23. Burfisher, M.E. & N.R. Horenstein, 1985. Sex Roles in the Nigerian TIV Farm Household. Women's
Role and Gender Differences in Development. Kumarian Press, Hartford, CT.
Feldstein, H. Sims & S.V. Poats (Eds), 1989. (Op. cit.).
24. Nagy, J.G., H.W. Ohm & S. Sawadogo, 1989. Burkina Faso: a Case study of the Purdue University
Farming Systems Project. In: H. Sims Feldstein & S.V. Poats (Eds) Working Together. (Op. cit.).
25. Paris, T.R., 1989. (Op. cit.).
26. Fernandez, M.E., 1994. Women's Agricultural Production Committees and the Participative-
Research-Action Approach. In: H. Sims Feldstein & J. Jiggins (Eds) Tools For the Field. (Op. cit.).
27. Poats, S.V., H.S. Feldstein & D. Rocheleau, 1989. Gender and Intra-Household Analysis in On-Farm
Research and Experimentation. In: R. Wilk (Ed) The Household Economy, Reconsidering the Domestic
Mode of Production. Westview Press, Boulder, CO.
28. Baker, D., 1994. Women and Trials Management in Botswana: Experiences with Farmer Groups. In: H.
Sims Feldstein & J. Jiggins (Eds) Tools For the Field: Methodologies Handbook for Gender Analysis in
Agriculture. Kumarian Press, West Hartford, CT.
29. Sutherland, A., 1994. Incorporating gender into a national programme of FSR. In: Feldstein, H. Sims & J.
Jiggens (Eds) Tools for the Field. (Op. cit.).
30. Ashby, J,. 1989. (Op. cit.).
31. Flora, C. Butler, 1994. Using focus groups with rural women. In: H. Sims Feldstein & J. Jiggins (Eds) Tools
For the Field. (Op. cit.).
Rocheleau, D.E., 1994. Investigating contradictions and mysteries. In: Feldstein & Jiggens (Op. cit.).
32. Thomas-Slayter, B., A. Lee Esser & M. Dale Shields, 1993. (Op. cit.).
Lightfoot, C., S. Feldman & M. Zainul Abedin, 1991. Households, Agroecosystems and Rural
Resources Management. Manila: Bangladesh Agricultural Research Institute and International Center for
Living Aquatic Resources Management.

76 Chapter 3

33. Narayan, D. & L. Srinivasan, 1995. Participatory Tool Kit. The World Bank, Washington, DC.
34. Buenavista, G. & C. Butler Flora, 1993. Surviving Natural Resource Decline: Explaining Intersections
of Class, Gender and Social Networks in Agbanga, Leyte, Philippines. An ECOGEN Case Study. Virginia
Polytechnic Institute and State University, Blacksburg, VA.
Thomas-Slayter, B., A. Lee Esser & M. Dale Shields, 1993. (Op. cit.).
35. Guijt, I., 1995. Questions of Difference: PRA. Gender, and Environment. A Training Video. International
Institute for Environment and Development, London.
36. Sperling, L., M. Loevinsohn & B. Ntabomvura, 1993. Rethinking the farmer's role in plant breeding:
local bean experts and on-station selection in Rwanda. Experimental Agriculture. 29, 509-19.
37. Rassam, A., 1985. Farm labor by age and sex in north-western Syria: implications for two proposed tech-
nologies. Paper presented at the Farming Systems Symposium, Kansas State University, Manhattan,
Kansas, 13-16 October, 1985.
Nour, M.A., 1985. In The Rockefeller Foundation and International Service for National Agricultural
Research. Women and Agricultural Technology: Relevance for Research. Volume 2: Experiencies in
International and National Research. Report for the CGIAR Inter-Center Seminar on Women and
Agricultural Technology. Bellagio, Italy. 25-29 March, 1985. The Hague. Netherlands.
38. Chabala, C. & R. Nguiru, 1989. Intrahousehold Dynamics and FSR/E in Zambia, a Case Study of
Traditional Recommendation Domain 3 in Central Province. In Working Together (Op. cit.).
39. Kennedy, E. & P. Peters, 1992. Household food security and child nutrition: the interaction of income
and gender of household head. World Development, 20(8), 1077-85.
Hopkins, I., C. Levin & L. Haddad, 1994. Women's income and expenditure patterns: gender or flow?
Evidence from Niger. Paper prepared for 1994 Annual Meetings of the American Agricultural Economics
Association, August.
Kennedy, E. & L. Haddad, 1994. Are pre-schoolers from female-headed households less malnourished?
A comparative analysis of results from Ghana and Kenya. Journal of Development Studies. 30(3).
40. Kennedy, E. & L. Haddad, 1994. Are pre-schoolers from female-headed households less malnourished?
A comparative analysis of results from Ghana and Kenya. Journal of Development Studies, 30( 3).
41. Paris, T.R., C.P. Diaz, M. Hossain & A.B. Vasallo, 1994. The process of technology development
and transfer to women: a case of the microrice mill in Guimba, Nueva Ecija. Philippines. Unpublished
42. Goetz, A.M., 1996. Local heroes: patterns of field worker discretion in implementing GAD (Gender and
Development) Policy in Bangladesh. IDS Discussion Paper 358. Institute for Development Studies.

Robert Tripp

One of the values of causal analysis in FSR is that it forces participants to take a step back, consider a wider
range of options than their own disciplines might suggest, and then focus on interventions that are rele-
vant to the given circumstances.

3.5.1 Introduction tion that technological innovations need to be
developed and tested under farming conditions
It is difficult to provide any precise definition and management representative of target
for a movement as broad as FSR, but two ele- farmers. It is the expectation of FSR that these
ments are certainly characteristic. One is an two elements should be directly linked: an
insistence that priorities for research and understanding of farming system performance
extension programmes must be based on an and constraints should naturally lead to the
understanding of farming practices within a identification of appropriate technologies to be
holistic framework. The second is the convic- tested on farm.

FSR Understanding Farming Systems 77

But the translation of farming system diag-
nosis to relevant technology testing has not
been straightforward. One technique has been
used to make a more effective connection
between diagnosis and experimentation the
management of causal analysis in FSR plan-
ning. This is a brief overview of the techniques
that are used, providing an evaluation of the
strengths and limitations of these techniques,
and placing causal analysis in the broader con-
text of the other adaptive and participatory
agricultural research activities currently in use.

3.5.2 The conduct of causal analysis
Most FSR has been directed at public sector
agricultural research and extension institu-
tions. The aim has been to provide methods and
techniques that are replicable and that can be
incorporated into the operating procedures of
these institutions. But the challenge has not
been merely methodological. A concomitant
purpose has been to impress upon researchers
and extensionists the complexity and rational-
ity of local farming systems and to convince
them that textbook solutions to farmers' prob-
lems are often inappropriate. One of the most
difficult aspects of conducting FSR, and in man-
aging related training activities, has been pre-
cisely the point at which a diagnosis of the
farming system is to be used to identify possible
Techniques for diagnosis, including various
types of surveys, interviews, group meetings
and observations, are relatively well described.
Similarly, a range of sources provide guidance
for the design of on-farm experimentation. But
there are no standard methods for connecting
diagnostic information to experimental design.
In the absence of a robust linkage process, one
unfortunate tendency has been to follow FSR
diagnosis with proposals for testing and pro-
moting the very technologies to which
researchers were already committed before the
This challenge of stimulating an innovative
approach to experimentation, consistent with
the insights of farming systems analysis, was
the principal motivation for the use of causal
analysis in FSR. Causal analysis is by no means
a universal component of FSR, however. The
description in this section is based in particular

on the experience of the CIMMYT and the
International Center for Tropical Agriculture
(CIAT), which used causal analysis in their
training activities in FSR1. Causal analysis is
part of a simple analytical sequence that
* An identification of problems that restrict
the productivity of a farming system.
* An analysis of the causes of those problems.
* Proposals for possible solutions based on an
understanding of the causes.
The process is best illustrated with the example
cited by Krisdiana et al. in 19912. An adaptive
research programme in Indonesia managed by
the Malang Research Institute for Food Crops
(MARIF) found that one problem in local maize
production was uneven plant populations with
high interplant competition. The immediate
cause of this problem was the high planting
density used by farmers. An initial response
might have been a recommendation to change
planting practices, but a farming systems per-
spective indicates the value of looking for the
rationale (i.e. the causes) behind the practice.
Three hypotheses were proposed (Fig. 3.5.1).
* Farmers may have overplanted as a reaction
to seed quality problems.
* Farmers may have placed priority on maize
thinnings as a source of animal feed.
* Farmers may have been trying to counteract
the effects of early season pest damage.
The identification of effective and thus appro-
priate solutions obviously depends on assessing
the relevance of these causal hypotheses.
Possible solutions for seed quality, such as modi-
fying sources, treatment or storage of maize
seed, would be very different from those
addressing the constraints of animal feeding,
such as the identification of alternative fodder
sources. In this example, the third causal
hypothesis was the relevant one, and further
research uncovered considerable shootfly dam-
age that had led farmers to overcompensate in
their planting practices.
The example illustrates several aspects of
causal analysis. First, chains of causes are
common; the immediate cause of the problem
was overplanting, but this in turn was suscepti-
ble to causal analysis. Second, causes are not
always immediately clear, and hypotheses need

Chapter 3

Poor seed quality and uncertain Farmers Use Very Farmers overplant to compensate for
germination rates [rejected] High Plant expected early season pest damage

Maize thinnings used as fodder
for livestock [rejected]


Fig. 3.5.1. Hypotheses on the causes of interplant
(Source: Krisdiana etal., 1991.)

to be investigated through further conversa-
tions with farmers, observations, surveys or
experiments. In addition, several layers of
causes may have to be investigated before arriv-
ing at one that is appropriate for suggesting
interventions. Once it was understood that
shootfly should be addressed, simple seed treat-
ments were tested. Further causal analysis
examined delays in planting and reliance on
maize-maize rotations as contributors to the
shootfly problem, with the aim of possible mod-
ifications in rotation or planting date to deal
with the pest problem (Fig. 3.5.2).
The complexity of causal analysis is such
that diagramming is often helpful, as illustrated
in Figs 3.5.1 and 3.5.2. This complexity is a
serious concern; there are, after all, almost no
limits to the breadth and depth of causal
chains. The technique is only useful if it is man-
aged as a pragmatic tool forcing researchers to
understand as much as they can about the con-
text of particular problems that have been iden-
tified before they start proposing interventions.
The causes of production problems can
include natural conditions (e.g. soil type, rain-
fall pattern), socioeconomic conditions (e.g.
food preferences, market requirements, current
policies) and management practices. The use of

competition in maize, Malang District, Indonesia.

non-practices as causes (e.g. farmers' unfamil-
iarity with chemical pest control in Fig. 3.5.2)
should be limited, otherwise causal analysis
becomes simply an inventory of the technolo-
gies that researchers or extensionists wish to
recommend without considering a systematic
diagnosis. The breadth of causal analysis
should correspond to the capacities and man-
date of the FSR programme. One common
debate, for example, is the degree to which pol-
icy factors should be included in a causal analy-
sis. To the extent that policies are susceptible to
the actions and recommendations of the FSR
programme, they are appropriately considered
in a causal analysis.
The examples discussed above involve single
problems, but the combination of the causal
analysis of several problems often reveals inter-
relations that help to identify further priorities
for research or intervention. It is not uncom-
mon to find that several causal factors con-
tribute to a particular problem and that they
must be considered together in testing possible
Causal analysis should not be confused with
two other common techniques in farming sys-
tems analysis3. It is not the same as problem
ranking, carried out after an initial set of prob-

FSR Understanding Farming Systems

lems have been identified and involving prioriti-
zation on the basis of such factors as the impor-
tance of the problems and their susceptibility to
amelioration. It is not the equivalent to analysis
of solution feasibility, where possible interven-
tions are ranked on the basis of parameters
such as cost of research. Causal analysis stands
in between problem identification and solution
It is also important to emphasize that the
sequence of problem-cause-solution should
not be confounded with a methodological
sequence in FSR. Although it is helpful to pre-
sent FSR as a progression from diagnosis to
planning to experimentation to assessment and
feedback, the consideration of problems, causes
and solutions goes on continually during the
conduct of FSR. Causal analysis should not be
seen as an isolated planning procedure but
rather as a guiding principle for the conversa-
tions, surveys, experiments and data analysis of
an entire FSR programme. The process of iden-
tifying and interrelating problems and causes
begins with the first activities in an FSR pro-

Delays in planting post-rainy
season maize because of upland
rice planted in rainy season

gramme and continues over each season with
further diagnosis and experimentation.
Hypotheses about problems and causes are dis-
cussed, tested and refined; the understanding of
these relationships should progress in parallel
with advances towards identifying useful inter-
ventions for the farming system.

3.5.3 Contributions of causal analysis
One of the values of causal analysis in FSR is
that it forces participants to take a step back,
consider a wider range of options than their
own disciplines might suggest, and then focus
on interventions that are relevant to the given
circumstances. In the Indonesian example,
extensionists might have recommended an edu-
cational programme to encourage lower plant-
ing densities; post-harvest specialists might
have embarked on a seed storage campaign;
socioeconomists may have encouraged more
attention to livestock management in the farm-
ing system. In the context of the immediate
problem and its actual cause, all of these would

Farmers unaccustomed
to the use of a chemical
pest control on maize

Late planting date for some
post-rainy season maize fields

Excessive turnaround time
between crop seasons

Delays in land preparation in
some fields due to excess rains


Fig. 3.5.2. Hypotheses on the causes of pest damage in maize, Malang District, Indonesia. (Source:
Krisdiana et al., 1991.)

80 Chapter 3

have been misguided. The search for causes is
an implicit part of most problem-solving tech-
niques, but finding ways to make it an explicit
and continuous part of FSR analysis stimulates
participants to seek more information and to
think carefully about the precise points of the
farming system where innovation would be
most productive.
It is important to realize that until the causes
of a problem have been identified it is often diffi-
cult to proceed with interventions. One of the
contributions of causal analysis to FSR plan-
ning is to encourage participants to articulate
whether they are designing research and exper-
imentation to test solutions for problems that
are understood, or to seek further information
to clarify what direction should be taken. In the
Indonesian example, research was done on the
quality of farm-stored maize seed, not as an
intervention but rather to test the possible role
of seed quality in the causal sequence.
Causal analysis helps provide a structure for
FSR planning that limits the tendency to pro-
mote preconceived 'solutions looking for prob-
lems'. It offers a framework in which to discuss
the rationale and interrelations of the farming
system that can be reviewed and refined. As
long as causal analysis is managed in a prag-
matic fashion, it helps researchers to focus on
creative solutions and to appreciate the degree
of understanding required to identify useful
innovations. Not enough attention to causes
leads to inappropriate research and recommen-
dations, but excessive pursuit of causal rela-
tionships diverts resources from the adaptive
aims of FSR.

3.5.4 The limitations of causal analysis
Causal analysis has proven to be a useful tool
for FSR planning, and for helping to communi-
cate the basic premises of FSR. Its use does not,
however, guarantee relevant research. As with
any aspect of research methodology, warnings
about the dangers of mechanical and unimagi-
native application hold true for causal analysis.
The identification of useful interventions in FSR
depends more on the skill and the experience of
the participants than on any planning tech-
nique. Causal analysis helps to keep disciplinary
bias at bay, but it certainly does not eliminate
the possibility that the most influential mem-

bers of the team will commandeer the direction
of an FSR programme.
One concern with techniques such as causal
analysis is that they may contribute to an exces-
sively positivist conception of FSR. The
sequence of problems, causes and solutions
gives the image of a completely objective project
aimed at discovering scientific truth. In fact.
adaptive agricultural research is a much more
subjective, iterative and political endeavour
than is often acknowledged. Any tendencies to
place such techniques on a pedestal of 'pure'
science should be counteracted. One way of
doing this is to admit that it is simply one more
technique to increase the chances that the
farmers, researchers and extension agents
brought together for FSR will interact with tol-
erance, creativity and understanding.
Indeed, it is worth briefly deflating the posi-
tion of causality in scientific endeavours. Science
is sometimes envisaged as the search for single
determining causes, but the reality of multiple
causation and the influence of a researcher's
frame of reference, and the scope of the enquiry
being conducted, challenge such simple concep-
tions4. The structure of causality is not objec-
tively determined, but is rather dependent on the
context of the research and the purposes of the
human actors5. In social science analysis,
causality is only one of several ways of describ-
ing relationships among variables6.
Another possible limitation to causal analy-
sis (as described here) is the fact that it was con-
ceived primarily as a technique for improving
the FSR planning procedures used by public sec-
tor researchers and extension agents. Although
FSR involves considerable interaction with, and
participation from, farmers, the use of causal
analysis in the planning of an experimental
programme does not envision significant direct
farmer involvement. As current movements in
adaptive agricultural research include consider-
ably more emphasis on farmer participation,
and often less of a presence of public sector
technical staff, the broader relevance of this
technique is worth exploring.

3.5.5 Causal analysis in a broader context
Various types of causal analysis are common in
project planning and design. Delp et al., in
1977, described several planning techniques

FSR Understanding Farming Systems

that involve causal analysis7. A number of
donors have adopted the GTZ 'objective oriented
project planning method', ZOPP7, which begins
by identifying a core problem and then analyses
its causes and effects as a prelude to specifying
possible alternatives for project attention.
Examples of the use of causal analysis in agri-
cultural research design include the method of
farming systems diagnosis described by
Lightfoot et al. in 19907 for training researchers.
It uses systems diagrams, where farmers' prob-
lems are placed in the centre and causes,
divided into primary, secondary, biophysical and
socioeconomic causes, are arranged in circles
around the problem.
The conventional concept of FSR planning is
that of researchers and extension agents in an
office or meeting room, debating the diagnostic
and experimental data available to them. There
is a need to involve farmers more in the plan-
ning process, and to consider the implications
for causal analysis. In a technique described by
Bunch in 1982 for World Neighbors activities8,
farmers first brainstorm to produce an initial
list of problems, then work to refine the list,
group similar concepts together and arrange
them in order of priority. The iterative nature of
planning would seem to indicate significant
possibilities for shifting the locus of much plan-
ning to the field, taking advantage of innova-
tions such as the 'regular research field
hearings' described by Baker in 19889 or the
farmer groups described by Norman et al.10,
that same year.
It is certainly the case that recent interest in
participatory agricultural research methods
provides few examples of what could be
described as causal analysis. Perhaps this
should not be surprising. First, there is a strong
reaction by many against the 'empiricism' of
FSR and indeed against what Roades termed
'the bankruptcy of social science methods used
by FSR teams' in 1994". The feeling here is
that less structure and a wider range of meth-
ods would be helpful. In addition, farmer partic-
ipation often emphasizes the importance of
indigenous knowledge and explanatory con-
cepts that may not be compatible with conven-
tional causal analysis. Finally, the importance of
empowerment12 in farmer participatory meth-
ods directs attention away from the details of
formal planning and towards the importance of

farmers assuming control and responsibility for
technology generation.
There is some debate regarding the degree to
which FPR should be used to refine FSR13, or
whether it should complement14, or indeed
replace it1 5. In any case, it can be argued that the
role of causal analysis in FSR can be usefully
considered for FPR as well. Work in FPR has
been responsible for a considerable expansion in
the range of field techniques available for diag-
nosis, and impressive lists of alternative methods
for engaging farmer experience and commit-
ment for problem identification and prioritiza-
tion are often presented16. However, this
methodological diversification has not solved the
problem of how to convert a description of con-
ditions and problems into a plan of work. As
Mosse said in 1996: 'The concept of "people's
knowledge" misrepresents information produc-
tion in the planning process and gives a decep-
tively participatory gloss to the more complex
social dynamics of knowledge and the process of
negotiation involved. ... Even where sophisticated
methods of participatory appraisal are sensi-
tively and effectively used, the local knowledge
which they help to generate does not in any
straightforward way translate into programme
decision-making and action'17. The richness of
participatory appraisal has yet to be matched by
an adequate strategy for participatory planning.
In this sense, FPR shares much in common
with FSR. Both have an unfortunate tendency
to focus excessively on their diagnostic tech-
niques; neither are very certain about how to
proceed from diagnosis to planning; and neither
can guarantee that their most powerful partici-
pants will not appropriate their course.
Techniques such as causal analysis make a
modest contribution to addressing these defi-
ciencies in FSR. Political control is not going to
be countered by methodological innovation.
But planning and priority setting in a frame-
work which encourages participants to exam-
ine their premises, acknowledge their biases
and commit themselves to exploring alternative
explanations can contribute to a more open and
responsive programme of technology genera-
tion. Causal analysis is simply one way of pro-
moting a state of mind in which puzzlement
and respect for the complexities of local farming
systems encourage an understanding that leads
to meaningful improvement.

Chapter 3

1. Tripp, R. & J. Woolley, 1989. The Planning Stage of On-Farm Research. Identifying Factors for
Experimentation. CIMMYT and CIAT, Mexico, D.E and Cali.
2. Krisdiana, R., M. Dahlan, Herianto, C. van Santen & L. Harrington, 1991. From diagnosis to
farmer adoption: MARIF's maize on-farm research program in East Java, Indonesia. In: R.Tripp (Ed.)
Planned Change in Farming Systems. John Wiley, Chichester.
3. Woolley, J. & R. Tripp, 1994. Priority setting tools for research groups. In: J.D. Meindertsma (Ed.) Setting
Research Priorities. Royal Tropical Institute, Amsterdam.
4. Susser, M., 1973. Causal Thinking in the Health Sciences. Oxford University Press, New York.
5. Humphreys, P., 1989. The Chances of Explanation. Princeton University Press, Princeton.
6. Rosenburg, M., 1968. The Logic of Survey Analysis. Basic Books, New York.
7. Delp, P., A. Thesen, J. Motiwalla & N. Shesadri, 1977. Systems Tools for Project Planning. Program of
Advanced Studies in Institution Building and Technical Assistance Methodology, Bloomington. Indiana.
GTZ, 1988. ZOPP: an Introduction to the Method. GTZ, Eschborn, Germany.
Lightfoot, C., V.P. Singh, T. Paris, P. Mishra & A. Salmon, 1990. Training Resource Book for Farming
Systems Diagnosis. IRRI, Manila.
8. Bunch, R., 1995. Two Ears of Corn: a Guide to People Centered Agricultural Improvement. World
Neighbors, Oklahoma City.
9. Baker, D., 1993. Inability of Farming Systems Research to deal with Agricultural Policy. Journal of
Farming Systems Research-Extension, 4(1), 67-86.
10. Norman, D., D. Baker, G. Heinrich & E Worman, 1988. Technology development and farmer groups:
experience from Botswana. Experimental Agriculture. 24, 321-31.
11. Rhoades, R., 1994. Farmer participation in priority setting. In: J.D. Meindertsma (Ed.) Setting Research
Priorities. Royal Tropical Institute, Amsterdam.
12. Okali, C., J. Sumberg & J. Farrington, 1994. Farmer Participatory Research: Rhetoric and Reality. IT
Publications, London.
13. Ashby, J. & L. Sperling, 1994. Institutionalizing participatory, client driven research and technology
development in agriculture. Agricultural Administration (Research and Extension) Network Paper no. 49.
Overseas Development Institute, London.
14. Okali, C., J. Sumberg & J. Farrington, 1994. (Op. cit.).
15. Pretty, J., 1995. Regenerating Agriculture. Earthscan, London.
16. (Ibid.).
17. Mosse, D., 1996. Local Institutions and Farming Systems Development: Thoughts from a Project in Tribal
Western India. Agricultural Research and Extension Network Paper no. 64. ODI, London.

Part II

The Applications of Farming Systems


Mike Collinson

The introduction to this book adopted a definition of farming systems research (FSR) as 'under-
standing farmers' livelihood systems'. I believe this narrow definition puts FSR in its proper place
without exaggerating its scope; as an 'aide' to research and development in traditional agriculture.
Applications of FSR use this understanding in technology choice and generation, in development
programming, and in policy formulation. I use the phrase development programming deliberately
to include a widening range of activities, including its application to the introduction of new enter-
prises, particularly cash crops for the market, and to adding value through both group action and
local processing. Farm system understanding offers a good basis for juxtaposing the cash, labour
and land demands of new development opportunities on to the resource endowments and require-
ments of the current system, identifying potential clashes in resource use and formulating strat-
egy for innovation. Understanding the farming system is also a potential asset for the successful
operation of most enabling organizations, for project development in storage, marketing, process-
ing and credit, as well as for capital works within the community.
Part II of this book describes the evolution of FSR applications. The emphasis is on technology
generation, the origin of FSR, and the term on-farm research (OFR) is used to embrace both FSR
and its application in on-farm experimentation (OFE), forming a useful shorthand when discussing
the full R & D process. Two examples of its application in extension are followed by a discussion of
its potential role in policy formulation.

The main contribution to Chapter 4 focuses on
the application of FSR to technology develop-
ment and was written by Ann Stroud and Roger
Kirkby who have over 40 years of African field
experience between them. They look at the his-
tory of OFE, review the evolution of concepts
and methods on the where, what, who and how
of OFE. They go on to assess the impact of FSR
on research process in three thematic research
areas; varietal improvement, agronomy and nat-
ural resource management, and livestock; they
also examine the impact of 25 years of FSR on
perspectives on technology development, and on
the roles of farmers, researchers and institu-
tions. Finally, they examine recurring issues in
OFR before offering their view of the future.
Chapter 4.2 is a case study of technology
development in action, focusing on a project

CAB International 2000. A History of Farming
Systems Research (ed. M. Collinson)

seen as fairly typical of the way in which OFR
has been implemented. Alistair Sutherland and
J.N. Kang'ara describe a UK-funded OFR project
in eastern Kenya, reaching some half a million
people in 70,000 households. Its operating
strategy was derived from an assessment of
local institutional circumstances, aiming to sus-
tain the processes it introduced after its comple-
tion. It showed results, in terms of benefits to
farmers, over a 4-year period.
Chapter 5 has three contributions. Cornelia
Flora and Charles Francis examine farming sys-
tems extension in the USA. drawing attention to
two cycles of FSR in extension in the USA, first
in the 1920s and 1930s when extension agents
first sought to address the farm as a whole and
latterly, since the 1980s with participatory
approaches growing stronger and farmers,
increasingly organized in groups, becoming


University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs