• TABLE OF CONTENTS
HIDE
 Front Cover
 Title Page
 Table of Contents
 Message from the president...
 From the editor's desk
 Tentative criteria for selection...
 On the systems dimension in FSR,...
 The conceptual basis for targeting...
 Adoption of soybean: A comparative...
 Impact of plan Sierra's outreach...
 An interrogative approach to system...
 Sustainability peri-urban vegetable...
 Agricultural production on the...
 Letter from the secretary/treasurer...
 Instructions to authors






Group Title: Journal for farming systems research-extension.
Title: Journal of farming systems research-extension
ALL VOLUMES CITATION PAGE IMAGE ZOOMABLE PAGE TEXT
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00071921/00012
 Material Information
Title: Journal of farming systems research-extension
Alternate Title: Journal for farming systems research-extension
Abbreviated Title: J. farming syst. res.-ext.
Physical Description: v. : ill. ; 23 cm.
Language: English
Creator: Association of Farming Systems Research-Extension
Publisher: Association of Farming Systems Research-Extension
Place of Publication: Tucson Ariz. USA
Publication Date: 1990-
 Subjects
Subject: Agricultural systems -- Periodicals -- Developing countries   ( lcsh )
Agricultural extension work -- Research -- Periodicals   ( lcsh )
Sustainable agriculture -- Periodicals -- Developing countries   ( lcsh )
Genre: periodical   ( marcgt )
 Notes
Dates or Sequential Designation: Vol. 1, no. 1-
General Note: Title varies slightly.
General Note: Title from cover.
General Note: Latest issue consulted: Vol. 1, no. 2, published in 1990.
Funding: Electronic resources created as part of a prototype UF Institutional Repository and Faculty Papers project by the University of Florida.
 Record Information
Bibliographic ID: UF00071921
Volume ID: VID00012
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 22044949
lccn - sn 90001812
issn - 1051-6786

Table of Contents
    Front Cover
        Front Cover
    Title Page
        Page i
        Page ii
    Table of Contents
        Page iii
    Message from the president of AFSRE
        Page iv
    From the editor's desk
        Page v
        Page vi
    Tentative criteria for selection of future articles for this journal
        Page vii
    On the systems dimension in FSR, by Richard Bawden
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
    The conceptual basis for targeting farming systems: Domains, zones, and typologies, by Keith M. Moore
        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
    Adoption of soybean: A comparative analysis of cultural practices in Zaire and Nigeria, by Dennis A. Shannon, Mwamba Kalala M., Kubengu Mudilamika and Mpoy Mudiamvita
        Page 39
        Page 40
        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
    Impact of plan Sierra's outreach programs on the sustainability of hillside farming in the Dominican Republic, by Scott G. Witter and Michael P. Robotham
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
    An interrogative approach to system diagnosis: An invitation to the dance, by P. G. Cox, A. D. Shulman, P. E. Ridge, M. A. Foale and A. L. Garside
        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
        Page 83
        Page 84
    Sustainability peri-urban vegetable production and natural resources management in Nepal, by Hans G. P. Jansen, David J. Midmore, and Durga Dutta Poudel
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        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
    Agricultural production on the extensive margin: The case of Mexico's northern Altiplano, by Paul N. Wilson and Gary D. Thompson
        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
    Letter from the secretary/treasurer of AFSRE
        Page 129
        Page 130
    Instructions to authors
        Page 131
Full Text
Volume 5, Number 2
1995


o u n a
for Farming Systems
Research- Extension








Journal
for Farming Systems
Research-Extension


Volume 5, Number 2, 1995


Published by
the Association for Farming Systems Research-Extension









Journal for Farming Systems Research-Extension
ISSN: 1051-6786
Editor
George H. Axinn
Department of Resource Development
Michigan State University
East Lansing, Michigan
U.S.A.

The Journal for Farming Systems Research-Extension is published by the
Association for Farming Systems Research-Extension (AFSRE), an
international society organized to promote the development and dissemination
of methods and results of participatory on-farm systems research and extension.
The objectives of such research are the development and adoption through
participation by farm household members of improved and appropriate
technologies and management strategies to meet the socioeconomic and
nutritional needs of farm families; to foster the efficient and sustainable use of
natural resources; and to contribute toward meeting global requirements for
food, feed, and fiber.
The purpose of the Journal is to present multidisciplinary reports of on-
farm research-extension work completed in the field, and discussions on
methodology and other issues of interest to farming systems practitioners,
administrators, and trainers. The Journal also serves as a proceedings for the
international Farming Systems Symposia from which selected and refereed
papers are included. It also welcomes contributed articles from members of
AFSRE who are unable to attend the symposia. Contributed articles will be
judged by the same review process as invited articles.


The AFSRE President is:
Dr. Nimal Ranaweera
Addl. Secretary (Project Development)
Ministry of Agric., Lands, & Forestry
"Sainpathpaya"
82, Rajamalwatta Road
Battaramulla
SRI LANKA
Fax +94-1-868919 or +94-8-88798
e-mail: Nimal@deptag.ac.lk


The AFSRE Secretary/Treasurer is:
Dr. Virginia Cardenas
Department of Agricultural
Education and Rural Studies
College of Agriculture
University of the Philippines at
Los Bafios, College, Laguna
PHILIPPINES'
Fax +63-2-813-5697 or +63-2-942-914
e-mail: VRC@mudspring.uplb.edu.ph


Correspondence regarding articles for this journal should be addressed to:
George H. Axinn, Editor, JFSRE, E-Mail -- axinn@pilot.msu.edu
FAX 517-353-8994 Postal: 313 Natural Resources Building, Michigan State
University, East Lansing, Michigan 48824-1222, U. S. A.


Journal for Farming Systems Research-Extension









Journal for Farming Systems Research-Extension
Volume 5, Number 2, 1995

CONTENTS

IV. Message from the President of AFSRE

V. From the Editor's Desk

1 On the Systems Dimension in FSR
Richard Bawden

19 The Conceptual Basis for Targeting Farming Systems: Domains, Zones,
and Typologies
Keith M. Moore

39 Adoption of Soybean: A Comparative Analysis of Cultural Practices in
Zaire and Nigeria
Dennis A. Shannon, Mwamba Kalala M., Kubengu Mudilamika and
Mpoy Mudiamvita

55 The Impact of Plan Sierra's Outreach Programs on the Sustainability of
Hillside Farming Systems in the Dominican Republic
Scott G. Witter and Michael P. Robotham

67 An Interrogative Approach to System Diagnosis: An Invitation to the
Dance
P.G. Cox, A.D. Shulman, P.E. Ridge, MA. Foale andA.L. Garside

85 Sustainable Peri-Urban Vegetable Production and Natural Resources
Management in Nepal
Hans G.P. Jansen, David J. Midmore, and Durga Dutta Poudel

109 Agricultural Production on the Extensive Margin: The Case of Mexico's
Northern A Itiplano
Paul N. Wilson and Gary D. Thompson

129 Letter from the Secretary/Treasurer of AFSRE


Vol 5, No. 2, 1995









MESSAGE FROM THE PRESIDENT OF AFSRE

Let me first and foremost wish you welcome to the latest issue of the Journal. It
has been a long time since the release of the last issue, and I apologize for the
delay in this volume being released.
There was a change in the editorship and the process of selection for a new
editor went through a protracted period. This took some time. However, I am
now pleased to announce that Dr. George H. Axinn of Michigan State
University, and a very old friend of the Association, has taken over the
editorship. He will continue for some time, initially handling Volume 5 and
Volume 6.
I am aware that a number of members were concerned about the lack of
publications, and even questioned the rationale of joining the Association.
While recognizing the delay, at no stage was the idea of abandoning the
Journal ever considered. Rather, it was the Association's considered view that
all attempts must be made to have the Journal continue, and make it
sustainable by canvassing for more and more high quality articles.
The rotational process of the Presidency has brought the current Presidency
to Asia, which I think is challenging to the Southern countries, and also will
provide a new dimension to the Association. The shift had its initial problems
of logistics, which I am pleased are now over. The Association's office has
now moved to the Philippines. The address of the Secretary/Treasurer, Dr.
Virginia Cardenas, is shown inside the front cover, and her statement to the
membership is at the end of this issue.
Once again, let me apologize for the delay and at the same time assure you
that the professional competence and the quality will continue.

Dr. Nimal Ranaweera
President









FROM THE EDITOR'S DESK


This issue of the Journal for Farming Systems Research-Extension has been a
long time coming. As mentioned by our President, Dr. Nimal Ranaweera, in
his letter at the beginning of this issue, Dr. Timothy R. Frankenberger shifted
to a different position over a year ago, and the Association for Farming Systems
Research-Extension was without an editor for many months.
As a former agricultural editor who has been working with farming systems
and other rural development related activities for almost fifty years in Africa,
Asia, Latin America, and the USA, I was facing the change from full-time
professor to working Professor Emeritus in 1996 with some trepidation. So
when a member of the AFSRE Board asked me if I would consider serving for a
while as Editor of JFSRE, I was enthusiastic. Now, after reviewing over
seventy manuscripts in the "backlog" for this journal, and assembling those
most nearly ready for this issue, my enthusiasm continues.
This issue opens with two conceptual papers which I believe reflect the
"state-of-the-art" in farming systems research-extension. Richard Bawden, in
the opening paper, draws attention to the crucial dimension which the word
systems reflects in the name of our Association, and the name of this journal.
His scholarly approach sheds light on the systemic essence of our approach to
research and extension about farming. In the second paper, Keith Moore also
points to the lack of conceptual precision and transdisciplinary consensus
among practitioners, and attempts to address this confusion by clarifying many
of the key ideas in the field.
The following two papers move directly to the farm. A group of four authors
working in Zaire and Nigeria focus on the soybean, and do a comparative
analysis of cultural practices in the two countries. Then, in the next paper, two
field workers in the Dominican Republic analyze sustainability of hillside
farming systems.
In the next article a group concerned with farming systems work in
Australia suggest a fresh new approach to thinking about the decisions which
farming people make every day
And this is followed by two more "production-focused papers. One
addresses vegetable production in Nepal, and the next describes and analyses
agricultural production in one part of Mexico.
All of these papers were in some state of review or revision when I came to
this assignment, and I am personally and professionally pleased to present them
in this issue. But as a newcomer to this assignment, I must admit that they
were selected because they seemed to fit what AFSRE is all about, and they
were also the most ready for publication among some seventy papers in various
stages of the process. Now I am in the process of working with peer reviewers
scattered all over the world, trying to sort out articles for the next two issues of
this journal. We shall also be having new papers presented at the 1996


VoL 5, No. 2, 1995









Symposium in Colombo in November, and I hope additional manuscripts will
be written by other members of the Association.
One of the most difficult tasks of the Editor, along with the reviewers, is to
determine the criteria for selection of new articles for this journal. After all,
there are hundreds of other journals in every aspect of farming, agriculture,
forestry, fisheries, livestock, and related fields in every part of the globe.
Which might better be published in a more focused disciplinary journal? And
which are ideally suited for the Journal for Farming Systems Research-
Extension? A goal for each issue might be papers from as many different parts
of the world.
After correspondence with officers of AFSRE, and some members of the
Board, the set of tentative criteria on the next page have emerged. They are
included here with a request for your input, as a reader of this Journal. Please
look at the list, suggest additional items if you wish, and suggest other changes
which should be made. Assuming that most articles might not meet all of these
suggested criteria, preference could be given to those which meet more than
other papers. I hope an Editorial Committee will soon be formed for this
Journal, perhaps with members from each continent. Your comments and
suggestions will be shared with this editorial committee, and perhaps we can
evolve a set of criteria which will be helpful to members of the Association,
readers ofthie Journal, and future editors. Please send me your ideas.

George H. Axinn
Editor, JFSRE




###

Important Notice

Please turn to the last page of this Journal for an important message from
our Secretary/Treasurer, Dr. Virginia Cardenas. For readers who are not
members of the Association for Farming Systems Research-Extension, and for
those who may not yet have renewed their membership, you will find
important information at the back of this issue.


Journal for Farming Systems Research-Extension









Tentative Criteria for Selection of Future Articles for this Journal

a. Articles will address farms as whole systems, including their
production and consumption, and involving the plants, livestock, and humans
on the farming system, or the institutional support services for farming systems,
including the policy structure.

b. If possible, field research would involve partnerships among
professional agricultural/forestry scientists, extension personnel, and farmers
(men, women, and children). Ideally, that includes collaboration in deciding
what to study, designing the research, collecting the data, analyzing and
evaluating the findings, etc.

c. Human data will be disaggregated by gender.

d. Where papers involve several layers of a systems hierarchy -- from
soil microenvironments (for example), to whole farms, to communities of
farms, to larger agro-ecological zones, to national policy -- they will deal with
these phenomena from a systemic perspective.

e. Papers which deal with FSRE as a concept, and its relationship to
other approaches to rural development, etc., will assume that FSRE refers to
work like that described in (a) through (d) above.

f. Papers should demonstrate the relationship between new material
being presented and the current literature in FSRE, as published in earlier
issues of this journal, proceedings of AFSRE symposia, etc.

In addition to Regular Articles, the Journal will occasionally include a
section of Field Notes. These are more brief, need not have literature citations,
but are expected to share relevant field experience which may be useful to other
practitioners and scholars.


VoL 5, No. 2, 1995









ON THE SYSTEMS DIMENSION
IN FSR


Richard Bawden'

INTRODUCTION

The trouble with Farming Systems Research (FSR), is the word in the middle.
While most of us are pretty clear when it comes to thinking about what
'farming' is, and about what we mean when we talk about 'research', the
notion of 'systems' is far more elusive. Farming is what farmers do, and
research is what researchers do. And farming research, is research into
farming. So where does the systems bit fit in?
This is not, by any means, a trivial question. The demands on agricultural
researchers are getting more and more complex, and it is therefore essential
that the theories that they hold, and the research and development methods that
they practice, are relevant to the situations being faced. The challenge of
agriculture of course, is to continue to produce more and better quality food to
feed an ever-growing global population, while trying, simultaneously, to
improve the welfare of farmers and the rural communities in which they live,
and the integrity of the global environment which all of us share. As there are
some disturbing signs, that none of these three goals are currently being
achieved, it is vital that research approaches to agriculture and rural
development, are subjected to critical review with respect to their relevance.
And this brings us back to this issue of the word 'systems' at the heart of
FSR, and what it is supposed to be signifying with respect to the nature of the
research and development which is conducted in its name. These matters are of
significance not just to the researchers themselves, but to policy makers and
donors, who have to take decisions with respect to the allocation of scarce R&D
resources, to educators, who have to be clear about the nature of the specific
research competencies that their students need to acquire, and above all, to
farmers everywhere.
The purpose of this paper.is to explore this matter of the systems dimension
of FSR, in a manner which, hopefully, will add to the clarifications which have
been sought about the 'true' nature of FSR, almost since the time of its
emergence as a major theme in international agricultural research. The
essential argument that I want to put, is that by paying more attention to this



1 Professor of Systemic Development, University of Western Sdydney
Hawkesbury, Richmond NSW Australia.


VoL 5, No. 2, 1995






BAWDEN


issue of 'systems', and what it might mean to both their theory and practice,
FSR practitioners could significantly improve the quality and relevance of their
approach to research. The systems sciences, which have evolved as the
sciences of complexity, have much to offer to agriculture and rural
development, and there is a lot to be gained from FSR practitioners learning
more about what it means to be systemic.
Attention to the systems dimension, would bring farming systems
researchers into closer affinities with other major systemic approaches to
agricultural and rural development including agroecosystems analysis and
development, systems analysis and simulation, and the emerging critical
learning systems approach to development.

THE SEARCH FOR CLARITY

It is generally accepted that the term Farming Systems Research emerged in the
mid- 1970s as applied to the development of technologies for small-scale
farmers of limited resources (Hildebrand 1982). In response to what was being
seen as inappropriate approaches to the technological needs of resource-poor
farmers in the Third World, the arguments presented in favour of FSR
supported two fundamental propositions: (i) "that development of relevant and
viable technology for small farmers must be grounded in a full knowledge of
the existing farm system, and (ii) that technology should be evaluated not solely
in terms of its technical performance, but in terms of its conformity to the
goals, needs and socio-economic circumstances of the targeted small farm
system, as well" (Merrill Sands 1986). These two new imperatives for research
revealed the farm management and agronomy 'pedigree' of the movement,
while also setting it clearly apart from the types of commodity and/or discipline
focused research which characterized the agendas of the international
agricultural research centres of the day.
Given that the initiatives for this new, more encompassing, research
approach to agricultural development came essentially from the 'ground-up',
and concurrently in various locations across the world, FSR evolved into more
of an umbrella term for a class of research approaches than a descriptor for a
particular research methodology. This was in clear contrast to other
approaches to agricultural development which also promoted the merits of
approaching whole farms as systems, such as agro-ecosystems analysis
(Conway 1985) or the earlier initiatives in agricultural systems analysis (Dent
and Anderson 1971). Under these diverse circumstances, it was not surprising
that a wide variety of different concepts, approaches and research methods
would be grouped under the FSR heading, leading to considerable confusion
about what the approach really meant.


Journal for Farming Systems Research-Extension






ON THE SYSTEMS DIMENSION IN FSR


Calls for greater clarity in the terminology of concepts and methods of FSR,
first emerged from major reviews of the enterprise, nearly two decades ago
(Dillon et al 1978; Gilbert et al 1980). Despite these appeals however, by the
mid-eighties, not only did confusion still exist, but the problem seemed to have
got more acute, as the range of activities encompassed by the term FSR, had
broadened considerably. This realisation triggered a flurry of responses, and a
number of writers sought to reduce some of the confusion by attempting to
clarify terms and concepts, and creating classifications of the approach. Fresco
(1984) for example, distinguished between Francophone and Anglophone
traditions, describing the former as a "more formal, long-term and large-scale
research undertaking aimed at developing the potential of a geographical
region" while the latter "does not aim at a profound change in traditional
agriculture, but rather at incremental changes". Others have suggested more
complex schemes. Simmonds (1985), for instance, suggested that one could
distinguish between (a) FSR in the strictest sense, (b) on-farm research, and (c)
new farming systems developments. Merrill Sands (1986), went further in
suggesting that six types were more appropriate, identifying these as (a)
farming systems analysis (FSA), (b) farming systems adaptive research
(FSAR), (c) farming systems component research (FSCR), (d) farming systems
base-line data analysis (FSBDA), (e) new farming systems development
(NFSD) and (f) farming systems research and agricultural development
(FSRAD).
The variations among these different FSR types of research activity are
associated with matters like the intentions of the researcher, the extent to which
farmers themselves are involved, the level of innovativeness, and the extent to
which researchers from disciplines beyond agriculture, are involved. In spite of
their variations, all of the approaches have in common, the fact (a) that they are
complementary to conventional commodity and disciplinary research, and (b)
that they are, to a greater or lesser degree, what Merrill Sands (1986) refers to
as "systems oriented": There are two aspects of this orientation, because of the
fact that both the object of FSR "is regarded as a system", and "because of the
interdisciplinary way in which it strives after problem solutions" (Brouwer and
Jansen 1989).
In other words FSR is 'systems oriented' because both the object to be
researched and the methods by which it is researched, can be regarded as
'systems' the farming system as an 'objective thing' on the one hand, and the
inquiry as a 'systems process'on the other. In practice I want to suggest that
rather than characterising most FSR endeavours, these two aspects, of research
object and research process, actually represent two different traditions: The first
could be seen as research into farming systems (which I shall refer to as first
generation FSR) and the second, systemsresearch into farming (or second
generation FSR). This distinction is very important, or it emphasises issues
concerned with the 'nature of nature' and 'the nature of knowledge' which sit


VoL 5, No. 2, 1995






BAWDEN


at the very centre of what we can call the 'systems movement'. They also
provide a focus for discussions about how the application of systems thinking
and practices can be improved to create a third generation FSR, which
combines both first and second, while adding new attributes of its own. This is
therefore a useful point at which to jump into the systems dimension of FSR,
with an elaboration of what currently seems to characterise it.

THE CURRENT SYSTEMS DIMENSION

A typical definition of a farming system from the 'systems orientation'
perspective above, is that offered by Shaner et al (1982): "a unique and
reasonably stable arrangement of farming enterprises that the household
manages according to well-defined practices in response to physical, biological
and socio-economic environments and in accordance with the household's
goals, preferences and resources. These factors combine and influence output
and production methods."
This definition provides us with an important insight into the matter of the
'systems dimension' of FSR, for it represents a prime example of what
Checkland (1988) refers to as the confusion in the use of the word 'system'
between "everyday language" and the "language of professional discourse".
There is little doubt that the statement above is much more reflective of
"everyday language" than of "professional discourse", with very little evidence
of any careful formulation of the system as an abstract coherent entity with
special properties. FSR practitioners typically talk of farming systems just as
they might talk casually about a nation's research system, or its health system,
or their institute's pension system. Thus, as emphasised elsewhere (Bawden
1991), the systems model in FSR rarely extends much beyond fairly loose
descriptions of sets of relationships between farming enterprises, the household,
and the environmental influences under which they both operate. In other
words, they really do not express the formality of systems conventions in terms
of 'systems organisation', 'structure', 'properties', 'boundary conditions',
'cybernetics' or 'behaviour', in any meaningful manner. Thus even the
statement that "the farming system is part of larger systems e.g. the local
community and can be divided into subsystems e.g. cropping systems"
(Shaner et al 1982), gives little evidence of the conceptual significance of such
a hierarchy, let alone the characteristics of the systems at each of the different
'levels'. Indeed the use of words like "larger" and "divided into", while quite
appropriate for "everyday language", are actually quite misleading in terms of
conveying the sense of the concept of the hierarchical relationships between
"subsystems", "systems" and "suprasystems", as they would be expressed in
"professional discourse". Similar comments can be made about the language
used by Fresco (1984) in the following statement about the concept of hierarchy


Journalfor Farming Systems Research-Extension






ON THE SYSTEMS DIMENSION IN FSR


in a cropping system: "at the lowest level, one finds the cell and the plant
organs, followed by the plant itself. Plants combine into crops, crops into fields
that may carry crop populations of various species and variety, weeds and
pathogens. The farm is situated at the next higher level. Groups of farms
combine into villages or land-use units. These in turn combine in regions,
which may cover a part of a country, an entire country or even a group of
countries"
In many ways this lack of conceptual rigour with regard to the perceived
nature of farming systems, is quite predictable and defensible, given the claim
that the foundations of FSR were practical rather than theoretical or
philosophical (Norman and Collinson 1985). And indeed there are exceptions:
Dillon (1984) for instance, presented a relatively formal model of a farm as a
"purposeful system" which included the following subsystems: (a) technical, (b)
formal structural, (c) psychological or informal structural, (d) goals and values,
and (e) managerial. There is little evidence however that this model of a
socio-technical system, which clearly reveals a farm management economics
perspective of a farming system, has been embraced by FSR practitioners, and
"every day language" continues to persist with regard to the systems dimension
of this, the first generation FSR movement research into farming systems.
This is not to state that other attempts have not been made to formalise the
systems dimension or the 'systemicity' of the approach as we might refer to it.
Some workers have developed interesting perspectives on farming systems
by integrating ecological principles with farm management ones in their
attempts to express the systems nature of their endeavours. Approaches such as
these, which reflect our second generation FSR through their emphasis on
systems research into farming, have allowed the incorporation of quantitative
mathematical tools such as computer simulations and optimisation techniques
(Hart 1982; Norman and Collinson 1985; Penning de Vries et al 1993). And as
has been illustrated by recent conferences on the application of systems
approaches to agricultural development (Jones and Street 1990; Penning de
Vries et al 1993) and by the emphasis in so many editions of the international
journal Agricultural Systems, there is certainly no shortage of base data for use
in both biological and socio-economic systems simulations. The sophistication
of so-called expert systems for agricultural application, are providing
considerable evidence in support of the contention that the second generation of
FSR is becoming increasingly robust, and a useful set of five categories has
been proposed to differentiate between the increasing variety of approaches in
this domain (Jones 1989). These range from "heuristic expert systems" which
come closest to approximating the sort of "seat of the pants" decision-making
strategies used by recognized experts, to "problem specific skills" which utilise
expert data bases.

"I*- .,


VoL 5, No. 2, 1995






BAWDEN


For all these aspects of the systems dimension of FSR, both 'first' and
'second' generation however, we are still left with the sense that the matter of
the application of formal systems concepts, theories and philosophies, remains
unaddressed within the approach. There are those who argue that farming
systems approaches to research are more systematic than systemic (Holt and
School 1985; Bawden 1990) more concerned with the rigour and linear logic
of the process, than with the systemic interconnections of either the object of
the research or the process used. It has even been suggested that "a critical
approach to study and to developing farming systems has hardly been
developed" with FSR "anything but a lucid concept" (Brouwer and Jansen
(1989).
With these provocative statements as our motivation, and the sections above
as the context, I must now turn to a critical elaboration of what I mean when I
talk of "being systemic" and the conceptual and methodological challenges that
that presents to FSR practitioners. In arguing the case for the systemic
transformation of FSR, I draw attention to the claim at the start of this paper
about the threefold complex challenges that we all face as researchers
concerned with sustainable agricultural and rural development. I want to
suggest that future improvements at any level of agricultural 'systems' must be
evaluated as much for its ethical defensibility as for its social desirability, as
much for its ecological responsibility as for its economic viability, and as much
for its aesthetic acceptability as for its technical feasibility. And I want to
further submit, that all of these aspects of 'development' must be regarded as
fundamentally achievable through the research and development methodologies
that we use. In this I am calling for a 'third generation FSR'; one which we
might call Critical Farming Systems Research, or perhaps more correctly,
Critical Systems Farming Research!

BEING SYSTEMIC

Whenever we approach issues in the name of something or another, we are
bringing to bear certain ways of thinking and acting, as well as certain values
and assumptions which characterise that way of being. When we are being
scientific, for instance, we approach the world through deductive or inductive
ways of thinking, and we test our thoughts by creating and testing hypotheses
about the world about us, in a very systematic, value-neutral and methodical
manner. In being scientific, we are also holding on to certain assumptions
about the nature of the world, including the humans within it, and about how
knowledge about both can be acquired. What we do in this world when we are
being scientific then, strongly reflects particular ways of thinking and particular
views of the world which we have learned, somehow and somewhere along the
line of our education and training as scientists. We rarely, if ever, think about
the composition of these worldviews, nor about the way the\ 'frame' the way


Journalfor Farming Systems Research-Extension






ON THE SYSTEMS DIMENSION IN FSR


we think about, or act in, the world. If we refer to these characteristics of
thinking and acting and assuming, as paradigms, we can submit that for most
of us most of the time, our paradigms remain tacit. I use the plural here,
because we do not necessarily think or act in the same way, or hold to the same
assumptions of beliefs, when we are being something other than being
scientific, like being superstitious, for instance, or being mystical, or being
parental! Thus each one of us can use different paradigms for different
occasions. There are even differences between particular paradigms within
science, as was first argued strongly by Thomas Kuhn (1970).
Following Burrell and Morgan (1979), and using now the language of
professional discourse, we can refer to four particular sets of assumptions in
order to discriminate between different paradigms:
* assumptions about the nature of reality (ontology)
* assumptions about the nature of knowing about reality
(epistemology)
* assumptions about ways ofinqyiry into the nature of reality
(methodology)
* assumptions about the way human beings are (human nature),
To rewrite our descriptions on being scientific in these terms, and choosing
the paradigm of science that a research agronomist would probably hold, we
can argue that when such a researcher approaches a technical problem he or
she assumes (believes) that:
* there is a reality 'out there' which exists independently of anyone
observing it (a realist ontology);
* objective, knowledge about that reality can be acquired as
scientific, value-neutral truth (an objectivist epistemology);
* explanations (hypotheses) about the nature of reality and
generated in response to objective observations in it, can be
validated through experimentation (an empirical methodology);
and
* humans are rational, objective, interest-seeking, goal-setting
beings (a rationalist interpretation of human nature).
It does not take much imagination to realise that a farmer brought up in a
culture in which magic imbues all of nature with a soul or 'animus' (animism),
with myth and legend the source of knowledge about it, holds very different
assumptions about the world and how we can come to know about it, to our
agronomist. But the differences do not have to be so extreme as this for there to
be very significant differences in paradigms, with a systemic paradigm for
instance, being very different from a non-systemic one.


Vol 5, No. 2, 1995






BAWDEN


It is now appropriate to turn our attention to a discussion of the way we are
when we are being systemic, and explore what this might mean through
reference to the four sets of assumptions above. As might be expected by now,
the situation is by no means straightforward, even if again we confine our
descriptions to the language of professional discourse; for the trouble is that
there are two essentially different schools of systemics which are often
differentiated by being termed 'hard' and soft' respectively. Let's start with the
'hard' case and explore the assumptions held by a 'hard' systems researcher,
for in most respects, he or she holds the same beliefs as the agronomist
described above, with one very notable exception.
It hasn't been mentioned yet, but there can be very important differences
within the same category of assumptions above. It is one such 'within-category'
difference that allows us to highlight the essential difference between a
conventional agronomist and a systems researcher. Thus, within a realist
ontology, there are two very different belief positions which we can refer to as
holism on the one hand, and reductionism on the other, with our systems
researcher holding to the former set of beliefs and our agronomist, the latter.
For our immediate purposes, holism refers to the belief that the world about
us is structured in the form of whole entities (which we will call systems) with
each system having properties different from the sum of its own parts, as well
as from other systems around it. All systems are part of other systems, just as
they themselves are composed of (sub)systems. Reductionism, on the other
hand presents the absolute opposite belief position: the world is not structured
into whole entities with special properties, but consists instead of a mass of
elements and events which may or not be causally related to each other. Any
whole entity that does seem to exist, will be equal to nothing more than the sum
of its parts, and if it should appear otherwise, that is merely a reflection of the
ignorance of the observer rather than any innate property of the entity itself.
Simple though this distinction between holism and reductionism might
sound, it represents very powerful differences in terms of paradigms, and this
means that the arguments in this paper for more attention to be paid to the
systems dimension of FSR, is actually a call for a profound shift in paradigms.
Let me add emphasis to this statement, by further exploring the nature of
systems and relating this to the notion of holism and to the concept of being
systemic.

THE NATURE OF SYSTEMS

The two notions that are central to holistic beliefs, and which therefore
shape the 'systems movement' are: that systems are coherent whole entities,
and that as coherent entities, they possess properties distinct from either their
component sub-systems or the suprasystems of which they themselves are part.
These unique characteristics are referred to as emergent properties as they


Journalfor Farming Systems Research-Extension






ON THE SYSTEMS DIMENSION IN FSR


emerge at each level of any systems hierarchy as a result of what we might refer
to as recursivee interrelationships' mutual relationships between subsystems
and systems and suprasystems in which each level of system influences the
others. From a holist perspective, surprise is anticipated, but never predictable.
No matter how comprehensive the studies of the relationships between
component subsystems, or between the system and its suprasystem are, they can
never reveal emergent properties.
As a whole entity, a system must have some way of staying whole. It must
therefore have a boundary which not only keeps it as a coherent whole, but
separates it from other systems and from its environment (which is itself of
course, a (supra)system). One way by which the bounded system retains its
coherence, is through the inter-relationships between its component sub-
systems, which because they are necessarily different from each other, are often
in tension with each other. It is the tension of these differences which gives rise
to the oneness and to the uniqueness of the entity. Thus as one of the pioneers
of systems thinking put it, "there is a glorious unity of difference" (Bertalanffy
1981).
The tension of difference that occurs within systems, is replicated between
them too; and especially between systems at different levels within systems
hierarchies. Systems need to have sufficient (requisite) variety and redundancy
to be able to deal with these inter-system tensions, just as they must have
sensitive mechanisms for communication and feedback (cybernetics) within
themselves, and between themselves and their environments. Another way of
putting this, is that systems must be flexible enough to deal with the changes in
their environment to which they are coupled through cybernetic processes. As
is being increasingly recognized with respect to the 'natural' environment
(Hollings 1995), these changes are often discontinuous and non-linear, and
therefore non-predictable. As chaos theory and systemic studies of complexity
are revealing, a small change in one system can often result in a very
significant (and most surprising) change in other systems far removed from it,
in either space or time, and occurring through processes of amplification that
probably can never be understood. This has enormous implications with
respect to the globalisationn' of 'local' effects such as pollution, or biocide
resistance, or micro-climatic change, or market dynamics. The opposite
situation can also occur, with similar comments to be made about the processes
of attenuation, in this case.
As this latter discussion reveals, systems, in addition to their wholeness,
and to their emergent properties, also have dynamics, and assumptions about
these are also of prime importance to researchers of systems. Matters of
systems dynamics are central to notions about their capacities for stability and
sustainability, and these in turn, are obviously of central importance to farming
systems and the nature of their relationships with other systems with which
they inter-relate.


VoL 5, No. 2, 1995






BAWDEN


There are many other characteristics of systems that are fundamental to
their nature, but the three we have just elaborated the coherence of
wholeness, the emergence of properties, and their dynamics are sufficient to
indicate the significance of renewing the focus on the systems dimension of
farming systems research.
So far we have been concentrating on concepts from the 'hard' systems
school, with its realist assumptions about the nature of reality. In other words
we have been talking about systems as they are presumed to exist in nature. As
has been pointed out elsewhere (Bawden et al 1985), inasmuch as FSR
practitioners are systemic in their research approaches to the development of
farming systems, they largely follow the traditions of the 'hard' systems school.
In other words, they approach cropping enterprises, whole farms, groups of
farms, or even entire rural communities, with a sense of their respective
wholeness' as if they believed that each level of organisation actually exists as a
'real' system within a hierarchy of systems, ranging from the cropping
enterprise right up to the community. Typically however, their interest in what
we might call formal systemics does not proceed much beyond this point.
There are very few studies indeed of the emergent properties of farming
systems for instance, or of their cybernetics, or of their 'tensions of difference',
or their requisite varieties or redundancies, or of systemic insights into the
inter-relationships of farming systems with their environments as suprasystems.
Principles of systems dynamics are rarely if ever invoked in debates about the
crucial issues of stability or sustainability of agricultural systems in first
generation FSR circumstances, although this is one of the potential
transformations that the second generation can bring and where of course, the
tools of systems simulation hold such promise.
Much could also be learnt from those involved in agroecological studies,
where formal systems concepts have been adopted, and used in the
measurement of flows of energy and material cycling through agricultural
ecosystems (Hart 1982). Of particular significance is the work of
agroecosystem analysis and its connection with development. Here
agroecosystems have been conceptualised as "well-defined systems of
cybernetic nature" (Conway 1987) and their properties explored specifically
from perspectives of their productivity, sustainablity and sustainability as well
as equitability and autonomy (Marten 1988). Conway (1990) has used these
concepts to make the point that agricultural development represents trade-offs
between systems properties, and it is to such matters that second generation
FSR practitioners are increasingly turning. Yet trying to model matters of
equitability and autonomy of ethics and social justice and symmetry of power
relationships as if they were objective and quantifiabe aspects of human
nature, is clearly not possible.
So it is important to now turn our attention to forms of being systemic
which move us beyond objectivism and allows us to explore situations using


Journalfor Farming Systems Research-Extension






ON THE SYSTEMS DIMENSION IN FSR


systemic concepts, but from a notion of 'human activity systems', and the 'soft'
school.


SOFT SYSTEMS PRAXIS

So far our systems focus has been on what we identified earlier as a perspective
which differed from conventional reductionistt) agricultural research only in
its holist assumptions, significant enough though these are. We now need to
take two more steps into the systemic world, which will move us very
significantly from this position into radically new territory yet territory
demanded by our need to focus on the sustainability of farming systems in
ways which include matters of ethics, aesthetics, and social justice as much as
productivity, economic viability and social desirability.
The first of these steps is to move from objectivist assumptions about how
things are known (epistemologies) to what we will refer to as relativist.
Drawing on Berstein (1983), we make this distinction in the following manner:
Objectivism means that there is some unchanging standard which we can use
as a framework to determine the nature of truth, and reality, and goodness or
rightness. Thus in our 'hard' farming systems above, nature is taken as an
objective reality about which scientific 'truth' can be discovered, and it is
'right' and 'good' to seek more productivity through the application of such
truths, in the name of progress. In contrast, relativism is the basic belief that all
of these concepts can only be as relative to some context or another such as
societal or cultural norms, or a particular set of theories.
The notion of 'systems' now takes on quite a different meaning from that
which we have been using so far: When we assume a relativist position on
holism, we are shifting from beliefs about how the world is in reality, to
plausible descriptions of it from different perspectives from system as
'objectively knowable thing', to system as 'an abstract concept'. In this manner
we can now talk about a system of inquiry as if it were a 'real' system with all
the characteristics of systems that we have described, while in fact, we know
that it is nothing but an abstract idea. Thus this is not the same as slipping
back into 'everyday language' for this time we hold on to the rigours of the
characteristics of systems as coherent wholes, which display emergence through
recursions, and which are dynamic in the face of environmental change.
Conversations between people can now be considered to be inquiry systems,
if they focus on how those people go about improving their own situations by
explicitly thinking and acting in systemic ways. In other words, sets of ideas
and communications can shaped in such ways that they become systems of
inquiry through the systemic actions of these involved in them! The concept of
wholeness through 'tension of difference' now takes on a very human face, as


Vol 5, No. 2, 1995






BAWDEN


the FSR practitioner faces up to the 'new reality' that different people can have
very different views indeed about the same situation. When stakeholders are
invited to participate in the development process and contribute to discussions
about what constitutes improvements, differences of opinion and conflict are
the norm rather than the exception.
To exemplify what all this means in FSR practice, let us turn to the vital
matter of citizen participation in the process of sustainable agricultural and
rural development, and about who decides what it is that constitutes
improvements to agricultural and rural situations, and what perspectives guide
them in their choice. Cornwall et al (1994), while remarking on the enormous
surge of interest in participation in the process of agricultural research and
extension, have also claimed that "most FSR/E scientists continue to investigate
for or sometimes on their farmer 'clients' rather than with them". In other
words, the farm continues to be regarded as a researched system rather than a
researching one; a system which is able to 'learn its way forward into better
futures' as a result of those who are participants in it, learning themselves to
think and act in systemic ways.
'Soft' systems thinking can liberate this situation in a systemic manner with
FSR scientists construing participation as essential to the way they think and
act systemically their systemic praxis, if you will for the process of
participation is a fundamental systemic issue. As Skowlimowski (1985) has it,
"wholeness means that that all parts belong together, and that means they
partake in each other. Thus from the central idea that all is connected, that each
is part of the whole, comes the idea that each participate in the whole. Thus
participation is an implicit aspect of wholeness".
There are of course many different degrees of participation, ranging from
mere tokenism to the true emancipation of citizens (Arnstein 1969), and this
matter will be an important one when we finally turn to the issue of criticality.
For now we can simply recognize that the higher the extent to which
participation occurs in the development process, the greater the range of
interpretations of what it is that constitutes improvements to specific situations.
Even within farming itself, it has long been recognized that there are a
multitude of purposes with which any farm can be endowed, with farmers
holding to complex sets of reasons for doing what it is that they do, and
involving instrumental, social, expressive, and intrinsic goals (Gasson 1973).
This purposeful nature of farming, presents such a complex mosaic for
discussions about what it is that constitutes improvements precisely because
people have such different values and worldviews from each other, and even
within themselves at different stages of their lives. Thus, as there will be
different notions of what constitutes truth, and justice, and fairness, and
goodness, and rightness, under such relativistic circumstances, it is vital that
the methods of systemic research include ways by which these differences can


Journal for Farming Systems Research-Extension






ON THE SYSTEMS DIMENSION IN FSR


be made explicit and accommodated in the search for agreement about
improvements.
The major insight of Peter Checkland, who is regarded as the 'father' of the
soft systems school, is that the exploration of any messy or complex situation
will reveal a number of different perspectives on that situation which can be
modelled as 'human activity systems', each of which reflects a specific
worldview, which therefore need to be made explicit through the process of
systemic inquiry. This need to qualify purposeful human activities by the
worldviews that support them in the search for improvements to situations by
participants in that situation, is very different from the ability to name a single
pursuit like productivity growth by an observer who is presumed to be
independent of that system. Herein then lies the essential distinction between
the 'soft' and the 'hard'. As Checkland (1995) has recently put it "the crucial
difference is between on the one hand an approach which assumes the world to
be a complex of systems, some of which may be malfunctioning, and on the
other an approach which makes no assumptions about the nature of the world,
beyond assuming it to be complex, but assumes that the process of enquiry can
be organised as a system of learning".
The methodology which Checkland developed from his systemic thinking
about messy, complex and purposeful situations, is thus itself approached as if
it were a learning system (Checkland 1981) with each of its stages being a sub-
system of the system of inquiry itself. In its use, it will reveal a host of issues of
significance to a range of stakeholders who are committed to coherent debates
about accommodating conflicting interests within political and social contexts,
and which enable action to be taken to improve problematic situations. As
Jackson (1995) has recently described this process, "the purpose is to generate a
systemic learning process in which the various participants in a learning
situation come to appreciate more fully, each other's world views and the
possibilities for change, and a consensus or at least accommodation (however
temporary) becomes possible between those who started with and may still hold
divergent views".
There are a number of key issues of importance to FSR practitioners, from
this 'soft' systemic perspective. The most obvious of these is that it provides a
systemic framework for the vital first step in FSR, of including the farmers
themselves in the process of decisions about what constitutes improvements. It
allows differences of opinion between researcher and farmer, and between
different farmers who might be participating together in a development project,
to be identified, and most crucially, linked to different worldviews, which are
made explicit. It also provides a vehicle for including other stakeholders in the
development process, and exploring the possible social, cultural and economic
impacts both positive and negative that changes at the level of individual
farms might have on the region as a whole. In this manner, some of the basic


VoL 5, No. 2, 1995






BAWDEN


concepts of systemic thinking, in terms of interrelationships, coherence,
hierarchy, etc are revealed and highlighted.
It follows from this that one of the key actions to be taken to improve the
problematic situation of FSR itself as a participative research approach, is the
improvement of the capacity of its practitioners to think and act systemically.
An important thesis in this regard, has been developed by Marcia Salner who
submits that individuals are only able to develop an effective systemic
capability once they have reached, what she refers to as a particular epistemicc
competence" associated with the acceptance of an epistemology of "contextual
relativism" (Salner 1986). In other words, what she is calling for, is equivalent
to a basic shift in one's personal assumptions from objectivismm' to 'relativism',
as we have defined these above. This is no light matter, for what she is
demanding is that we not only learn to question some of the most fundamental
beliefs that we hold, but that we are prepared, under certain circumstances, to
suspend and even change them to accommodate different circumstances.
Knowing what epistemological (and ontological) assumptions one is making
under any circumstance, is a critical first step in knowing the advantages of
also embracing other worldviews, and this means, in the present context that
FSR practitioners need to learn how to learn about being systemic (Bawden and
Packham 1993). Yet there is.a fundamental paradox here: FSR practitioners
need to reach a particular level of epistemicc' development before they are able
to really see the merits of thinking and acting in systemic ways. However they
are only likely to reach this stage of development if they are self critical of their
present approaches, and the assumptions upon which they are based They
need to be able to think in systemic ways in order to appreciate the advantages
of thinking in systemic ways!
And this brings us to the last of our areas of exploration; criticality.

THE ESSENCE OF CRITICALITY

Perhaps the easiest way of introducing the essence of being critical, is to focus
on the issue of assumptions about human nature that were introduced earlier. In
the discussion about being scientific, it was proposed that human beings were
rational in an objective way. Experience tells us however, that that is often not
the case, nor is it confined only to situations referring to an objective world for
there are other worlds too; thus Jurgen Habermas (1985) refers to "the social
world of legitimately regulated interpersonal relations", and each person's
"own subjective world of experience", in addition to an objective world. The
search for agreement about what needs to be done in the name of improvements
must therefore take account of all three of these worlds. For the FSR
practitioner, this means that the objective world of the 'hard farming system' is
set within the 'soft' relativist world of social difference, which itself is set
within interpretations posed by personal subjective experience.


Journal for Farming Systems Research-Extension






ON THE SYSTEMS DIMENSION IN FSR


According to Habermas, action oriented to reach shared understanding -
what he terms communicative action occurs through a common endeavour to
achieve consensus in situations where participants are free to state their views
and have an equal chance to do so (Habermas 1984). Unequal distributions of
power, lead to distortions in such communication (Habermas 1984) and thus
the ultimate goal of communicative action is one of emancipation.
A number of workers have taken these key ideas and incorporated them into
what they call 'critical systems' approaches, where the practitioners assume
positions of criticality of everything they think and do: from the
epistemological / ontological / methodological / ethical assumptions that they
and others in the situation hold, as well as of the contexts of power and
communication that prevail (eg Flood 1990; Jackson and Jackson 1991: Ulrich
1983). None of these endeavours have directly involved matters of agriculture
or rural development, although recent proposals by Ulrich (1996) to extend
critical systems thinking and actions to "citizens", is particularly pertinent to
the situation that we face in agriculture, given the fact that our constituency for
sustainable food production is essentially everyone in every community across
the globe!
There is no doubt that critical systemic approaches would be entirely
appropriate to agricultural and rural development situations, and represent
systems dimensions that need to be urgently explored in the next phase of
development of FSR as it moves into its third generation.

A SYSTEM OF FSR SYSTEMS

Following the logic established during the discussion about the 'soft' school, it
will now be apparent that what has been presented here as critical systemics,
soft systemics and hard systemics, itself represents a system of inquiry systems
which would be entirely appropriate for FSR. Each of these three 'levels' of
inquiry would reveal quite different emergent properties about situations in
which improvements were sought in the name of the sustainable agricultural
and rural development. And there will frequently be different interpretations
and conflicts both within each domain as well as between them. What seems
relatively straightforward as an improvement in the technical performance of a
'farming' system, is clearly much less so when considered from the point of
view of who might benefit from changes made, and who might be penalised,
and who decides which is which. And this in turn is further complicated by the
fact that discussions with participants about the 'best' strategies for
development, give a highly biased picture as a result of the distortions of
communication that are occurring within the community as a function of the
asymmetry of power relationships within it. Critical inquiry aims to clarify the
third of these matters, soft systemics are appropriate for exploring human
activities, and hard systems methodologies are needed to explore strategies for


Vol 5, No. 2, 1995






BAWDEN


change at the technical level. Clearly, each level of inquiy is critically
dependent on the other two.
As has been argued, the multi-facets of responsible development dictate the
need for FSR practitioners to consider the ethical, aesthetic, social, and
ecological aspects of their activities, in addition to those economic and
technical ones which are conventionally included. There is little hope of this
being achieved in the absence of a systemic perspective which portrays the
sense of wholeness in all of this and which allows for the participation of
relevant stakeholders in conversations which are as free from communicative
distortions as possible.
The case has been made elsewhere (Bawden and Packham 1993; Bawden
1995) that this in turn, demands particular forms of education and training
which allow epistemic capabilities relevant to systemic competence, to be
developed.
To those who would be dismayed by the theoretical / philosophical focus of
this article, I conclude with two observations: firstly that there is nothing so
practical as a good theory, and secondly, that any approach which claims to
embrace 'systems' in its title, needs to be able to defend that position.
In the face of the daunting challenges of the need for continuing growth of
global food production without compromise to environment or community, we
have little choice but to continually seek ways of improving our praxis as
responsible, ethically defensible and hopefully systemic practitioners.

REFERENCES

Amstein, S. 1969, Ladder of Citizen Participation. Journal ofAmerican Institute of Planners
35:216-224
Bawden, R.J. 1990, Of Agricultural Systems and Systems Agriculture In Systems Theory Applied
to Agriculture and the Food Chain (J.G.W. Jones and P.R.Street eds) Elsevier Appl. Sci. Publ.
New York.
Bawden, R.J. 1991, Systems Thinking and Practice in Agriculture. Journal of Dairy Science
74:2362-2373.
Bawden, R.J. 1995, Systemic Development: A Learning Approach to Change Occasional Paper #1.
Centre for Systemic Development. University of Westem Sydney Hawkesbury.
Bawden, R.J. and Ison, R.L. 1990, The Purpose of Field Crop Ecosystems: Social and Economic
Aspects In Field Crop Ecosystems of the World, J.G.Pearson (ed) Elsevier Londn Ch. 2. pp 11 -
35.
Bawden, R.J. and Packham, R.G. 1993 Systemic Praxis in the Education of the Agricultural Systems
Pracitioner. Systems Practice 6: 7-19.
Bawden, R.J., Iscn, R.L., Macadam, R.D., Packham, R.G. and Valentine, I. 1985, A Research
Paradigm for Systems Agriculture in Agricultural Systems Research for Developing Countries.
J.V. Remenyi (ed) ACIAR Canberra.
Bertalanffy, L. von 1981, Systems View of Man. (Paul Violette ed.) Westview Press. Boulder.
Bemstein, R.J. 1983 Beyond Objectivism and Relativism. Basil Blackwell Ltd. Oxford.
Brouwer, R. and Jansen, K. 1989, -Critical Introductory Notes on Farming Systems Research in
Developing Third World Agriculture. Systems Practice 2:379-387.
Burrell, W.G. and Morgan, G. 1979, Sociological Paradigms and Organisational Analysis.
Heinemann. London.


Journalfor Farming Systems Research-Extension








ON THE SYSTEMS DIMENSION IN FSR


Checkland, P.B. 1981, Systems Thinking: Systems Practice. John Wiley. Chicester.
Checkland, P.B. 1988, The Case for 'Holon'. Systems Practice 1: 235 238.
Checkland, P.B. 1995, Model Validation'm Soft Systems Practice. Systems Research 12: 47-55.
Conway, G.R. 1985, Agroecosystem Analysis. AgriculturalAdministration 20: 31-55.
Conway, G.R. 1990, Agroecosystems. In Systems Theory Applied to Agriculture and the Food Chain.
(J.G.W. Jones and P.R.Street eds) Elsevier Appl. Sci. Publ. New York.
Cornwall, A, Guijt, I., and Welboume, A 1994, Acknowledging Process:Challenges for Agricultural
Research and extension Methodology. in Beyond Farmer First. I. Scoones and J. Thompson (eds)
Intermediate Technology Publications. London.
Dent, J.B. and Anderson, J.R. 1971 (eds) Systems Analysis in Agricultural Management. Wiley.
Sydney.
Dillon, J.L. 1984, The Farm as a Purposeful System. Mimeograph Dept. Agric. Economics. University
of New England. Armidale.
Dillon, J.L., Plucknett, D.L., and Vallaeys, G.J. 1978, Farming Systems Research in the
International Agricultural Research Centres Rome Italy. Technical Advisory Committee to the
Consultative Group on International Agricultural Research.
Flood, R.L. 1990, Liberating Systems Theory. Human Relations. 42:216-22
Flood, R.L. and Jackson, M.C. 1991, Creative Problem Solving: Total Systems Intervention. Wiley.
Chichester.
Fresco, L. 1984, Comparing Anglophone and Francophone Approaches to Farming Systems Research
and Extension. Paper 4th Annual Farming Systems Research Conference. Kansas State
University.
Gasson, R. 1973, Goals and Values of Farmers. Journal ofAgricultural Economics 14: 521-538.
Habermas, J. 1984, The Theory of Communicative Action vol 1 Reason and the Rationalisation of
Society. Beacon Press. Boston.
Habermas, J. 1985, Questions and Counterquestions. in R.J. Bemstein (ed) Habermas and Modernity.
Blackwell. Cambridge.
Hart, R.D. 1982, The Ecological Conceptual Framework for Agricultural Research and Development.
in W. Shaner, P. Philipp, and W. Schmehl (eds) Readings in Farming Systems Research and
Development. Westview Press. Boulder.
Hildebrand, P. 1982, Motivating Farmers to Accept Change. in W. Shaner, P. Philipp, and W. Scmehl
(eds) Readings in Farming Systems Research and Development. Westview Press. Boulder.
Hollings, C.S. 1994 An Ecologist View of the Malthusian Conflict. in Population,
EconomicDevelopment, and the Environment (K. Lindam-Kiessling and H.Landberg, eds)
Oxford University Press. New York. Ch. 4 pp 79-103.
Holt, J.E. and D. School 1985, Technological Change in Agriculture the Systems Movement and
Power. Agric. Systems 28:69.
Jackson, M.C. 1995 Beyond the Fads: Systems Thinking for Managers. Systems Research 12: 25-42.
Jones, P. 1989, Agricultural Application of Expert Systems Concepts. Agric Systems 31:3.
Jones, J.G.W. and P.R. Street (eds) 1990, Systems Theory Applied to Agriculture and the Food
Chain. Elsevier Appl. Sci. Publ. New York.
Kuhn, T.S. 1970, The Structure ofScientific Revolutions. The University Press. Chicago.
Marten, G. 1986, Productivity, stability, sustainability, equitability and Autonomy as Properties of
Agroecosystem Analysis. Agric. Systems: 291- 316.
Norman, D. and Collinson, M. 1985, Farming Systems Research in Theory and Practice. in
Agricultural Systems Research for Developing Countries J.V. Remenyi (ed)ACIAR Canberra.
Penning de Vries, F., Teng, P. and Metselaar, K. 1993 (eds), Systems Approaches for Agricultural
Development. Kluwer Publishers. Dordrecht.
Salner, M. 1986, Adult Cognitive and Epistemological Development Systems Research 3: 225-232.
Shaner, W.W., Philipp, P.F. and Schmehl, W.R. 1982, Farming Systems Research and Development:
A Guideline for Developing Countries. Westview Press. Boulder.
Skowlimowski, H. 1985, The Co-operative Mind as a Partner of the Creative Evolution. Proceedings o0
the 1st Interational Conference on Mind-Matter Interactions. Universidada Estadual de Campinas,
Brazil.


VoL 5, No. 2, 1995







BAWDEN


Ulrich, W. 1993, Some Difficulties of Ecological Thinking, Considered from a Critical Systems
Perspective: A Plea for critical Holism. Systems Practice 6:583-611.
Ulrich, W. 1996 Critical Systems Thinking for Citizens: A Research Proposal.. Research Memorandum
#10 The Centre for Systems Studies. The University of Hull.


Journal for Farming Systems Research-Extension










THE CONCEPTUAL BASIS FOR
TARGETING
FARMING SYSTEMS:
DOMAINS, ZONES AND TYPOLOGIES

Keith M. Moore1



ABSTRACT


Farming system concepts have increased the capacity to
understand and provide assistance to farm households
throughout the world. Unfortunately, identifying and
resolving problems specific to particular types of farming
systems often suffers from a lack of conceptual precision and
transdisciplinary consensus among practitioners. This paper
addresses this conceptual confusion by clarifying farming
system concepts and their application in the targeting of
specific farming systems. It begins at the generic level of
concept formation by locating farming systems within the
context of an agricultural systems hierarchy, highlighting the
significance of the choices and limits inherent in the systems
approach. A second level of conceptual tools (domains, zones
and typologies) are then discussed. These concepts are
necessary for the targeting of specific groups of farm
households with common problems and for which common
solutions can be identified. At this level, systems logic must
also be applied as more precise specification of particular
research domains leads to the identification of homogeneous
types of farming systems. Such specification requires more
in-depth preliminary investigation than normally conducted
and greater emphasis on the socio-economic conditions
differentiating farm households.


1
Rural Sociologist, INRA/MIAC Aridoculture Center, Settat, Morocco. This paper was written as
part of the MIAC/INRA Morocco Dryland Agriculture Applied Research Project under USAID Project
No. 608-0136.


Vo. 5, No. 2, 1995






MOORE


INTRODUCTION

Targeting agricultural research and development for a farming clientele
requires a clear set of conceptual tools. Farming systems practitioners have
provided such tools. However, considerable confusion exists in the literature
and in practice about the application of farming system targeting concepts.
This confusion stems from two sources: (1) the haste to demonstrate impact;
and (2) the lack of a transdisciplinary theory of farming systems.
The applied nature of farming systems research and development
encourages a problem-oriented focus for practitioners. Farmer problems have
priority over reflection on conceptual issues. In practice, this orientation relies
heavily, if not consciously, on the disciplinary training of the
practitioners. Conceptual consensus is difficult to attain. It is easier to move
on to the next practical "problem". This reliance often violates the assumptions
of the farming systems approach, leading to mis-specification of both
problems and solutions.
Diagnosis of farmer problems functions on two distinct conceptual levels
and necessitates more preliminary investigative work than is usually
envisioned. The concept of farming system provides a means to understand
the dynamic totality of agricultural development. However, this way of
thinking cannot be applied directly to the solving of farmer problems.
Empirically-existing farming systems are too heterogeneous. A second level of
concepts is needed to focus efforts in the specification of different types of
farming systems for which specific problems can be identified and solutions
developed.
This paper clarifies the targeting concepts of farming system research
and development. Concepts are defined and their epistomological and
practical interrelationships specified. It then provides a checklist of criteria
for differentiating types of farming systems.


THE CONCEPT OF SYSTEM


Central to the farming systems research and development approach is the
concept of system (Shaner et al., 1982; Marcotte and Swanson, 1987; FAO,
1989). However, its familiarity as a concept within different disciplinary
traditions and applied contexts has lead to confusion within the literature and
between colleagues (Gibbs, 1985; Brossier, 1987). In order to clarify the role
that system plays, it is useful to recall the generic properties of systems: (1)
systems are holistic integration of interrelated elements; and (2) although set


Journal for Farming Systems Research-Extension






THE CONCEPTUAL BASIS FOR TARGETING FARMING SYSTEMS


within given boundaries, they are open to interaction with their
environments (Shaner, 1982; Conway, 1986; FAO, 1989).


The popularity of the systems concept among agricultural scientists
stems, pragmatically, from the desire to effectively grasp the dynamic totality
of agricultural development. The interrelationships' between the factors
which combine to generate agricultural products and sustain the lives of
those producing them are complex. They cannot be simply reduced to a few
causal relations between variables or a function of one of its components
(Baker and Norman, 1990). The whole is greater than the sum of its parts.
Agricultural production depends on the interaction of all physical,
biological and socio-economic factors (Shaner, 1982; Brossier, 1987). A
change in one element leads to changes in others, and the combined effect
reverberates throughout the whole. The concept of system provides a general
framework in which to account for this dynamism and complexity.
Agricultural systems, however, have been conceived at many levels.
These systems can be understood as hierarchically-nested in one another: (1)
from the physical and biological systems at the field level; (2) nested in crop
and livestock systems to a farm household system; (3) nested in community
and infrastructural systems composing the regional and national
agricultural systems. In fact, a multiplicity of agricultural system hierarchies
have been specified (e.g., Hart, 1982; Gibbs, 1985; and Conway, 1986). This
multiplicity has led to confusion between different approaches (Gibbs, 1985).
The system level (and, consequently, the hierarchy in which it is nested) that
becomes the object of study and intervention is often dependent on the
disciplinary background of the practitioner (Gibbs, 1985) and the
technological biases inherent in the agricultural development approach (Oasa
and Swanson, 1986). The open relationship between system levels and their
environments is the conceptual root of this problem.
In their review of farming systems literature, Brush and Turner (1987)
suggest two caveats to systems analysis that provide guidance in the
clarification of the systems approach.

First, systems must be artificially limited if analysis is
to proceed, and second, described systems are, therefore,
heuristic and artificial analytical devices rather than
natural phenomena. Although the systems approach tends
to push the investigator toward descriptive holism,
effective limits on the scope of system modeling are imposed
by problems of scale. The larger the descriptive model the
weaker it becomes for describing specific behavior.
Finer-grained models, on the other hand, may lack the


VoL 5, No. 2, 1995






MOORE


explanatory power of more general ones. It needs to be
emphasized that the systems approach is descriptive rather
than explanatory. It helps identify what processes exist
and how sets of interrelated components function together.
(Brush and Turner, 1987:27; emphasis in original).


The point is that systems are useful, albeit artificial, constructs. As such,
choices must be made to limit the scope of the work for heuristic purposes.
The consequent product is descriptive of targeted farming systems, rather than
construction of causal models or policy prescriptions (Whyte, 1983). It is
expected that policy decisions and causal modeling will be shaped by farming
system descriptions. Following sections describe the choices and conceptual
limits that are imposed in such a task.


THE FARMING SYSTEM AND ITS ENVIRONMENT

Since the core of confusion in farming systems research involves the level at
which the farming system is defined, it is useful to introduce a hierarchy of
agricultural systems. Hierarchical presentations are founded on the idea that
systems are nested in one another, each higher level often assumed to
encompass a larger geographical space or range of phenomena. This
simplifying assumption allows for greater facility in presenting relationships
between agricultural systems, but involves an explicit choice in how real world
conditions are to be conceived and described.
Two issues are involved. First, a progressive centralization of system
functions occurs in ascending a systems hierarchy. This results in
reductionist analyses (Marcotte and Swanson, 1987) or, simply, less capacity to
describe specific behavior (Brush and Turner, 1987). It also has important
implications for analysis since system functioning at one level (the farm
household) may be seriously affected by system components at another level
(national agricultural policy), yet the interrelationships have been obscured by
the nested hierarchy.
Second, there is no single hierarchy. Various agricultural system
hierarchies have been designed to focus on specific problem areas. The
traditional farming systems approach is conceived with reference to
productive units or specific enterprises within those units. Commodity system
approaches (e.g., Friedland et al., 1981) are conceived with reference to the
production, circulation and distribution of specific commodities.
Agro-ecological zones (e.g., Conway, 1986) and land-use patterns (e.g.,
ICARDA, 1989) have also provided the basis for hierarchical structures.
These latter are conceived with reference to the bio-physical resource base in


Journal for Farming Systems Research-Extension






THE CONCEPTUAL BASIS FOR TARGETING FARMING SYSTEMS


which agricultural production is conducted. Each of these highlights
certain interdependencies while others are left unaccounted.
The agricultural systems hierarchy presented in Figure 1 does not resolve
these issues. Rather, it provides an overarching frame of reference for the
farming systems approach presented here, as well as a means to integrate
analyses across system levels. The farm household is the chosen focal point
because the objective is to characterize the systemic conditions which
influence the adoption, adaptation or rejection of new agricultural technologies.
This focus assumes household units of production and consumption, where all
decisions on allocating human and material resources must ultimately be
reconciled. Norman and Gilbert (1982) provide a concise working definition of
a farming system.


The specific farming system adopted by a given farming
household results from its members, with their managerial
know-how, allocating the three factors of production (land,
labor, and capital) to three processes (crops, livestock, and
off-farm enterprises) in a manner which, with the knowledge
they possess, will maximize the attainment of their goalss.
(Norman and Gilbert, 1982:17)

This conception of a farming system is open to both the
socio-economic environment, usually considered in terms of higher system
levels, and the bio-physical environment. The bio-physical environment can be
broadly-defined in terms of agro-ecological zones at higher system levels
and more narrowly-defined in terms of biological and physical interactions at
lower levels. Lower system levels involving the management of crops,
livestock, and non-farm activities are developed as nested in the farming
system and referred to for purposes of clarity as sub-systems (see Figure 2).
Household resources of land, labor and capital, although allocated across
sub-systems as the wheels of production are set in motion, are kept analytically
distinct for conceptual clarity.
Conceptual models (Shaner et.al., 1982) or flow diagrams (Garrett, 1984)
are often presented with more detailed specification of sub-system components
and their linkages. This is an important way of suggesting interdisciplinary
data needs, and ultimately leads to the quantification of material flows
between components. Although many flow diagrams have been referred
to in the development of criteria for typology construction (below), such
specification must remain open until a particular research domain and type of
farming system are identified and described.


Vol 5, No. 2, 1995






MOORE


Figure 1: Hierarchy of Agricultural Systems


Journal for Farming Systems Research-Extension






THE CONCEPTUAL BASIS FOR TARGETING FARMING SYSTEMS


Cropping .- LiUvestock
Sub-System Sub-System


'arm
( Household
\ ....... I\J


Figure 2: The Farming System


TARGETING DOMAINS

Applying system concepts in agricultural development requires an
intermediary set of conceptual tools to provide a bridge from the theoretical
understanding of farming systems to the applied targeting of technology
research and development.
As with the system concepts, considerable confusion exists concerning
the specification of targeting concepts (Wotowiec et al., 1988). Common
practice within the farming systems research and extension tradition has been
to quickly survey a region and establish recommendation domains on the basis
of easily identifiable characteristics (Simmonds, 1985). Recommendation
domains are also specified at other stages of the technology research and
development process. Consequently, Wotowiec et al. (1988:77) note, the term
"'recommendation domain' has been stretched to cover too many situations
and too many different purposes." Such hastily conceived recommendation
domains are founded on the assumption of homogeneity among both farmers
and farming conditions. They further note that homogeneous groups
should not be identified until "researchers have an adequate understanding


VoL 5, No. 2, 1995






26 MOORE

of the variability inherent in local farming systems, usually not
accomplished early in the work in an area" (Wotowiec et al., 1988:74).
The following specification of targeting domain concepts is based on
Wotowiec et al. (1988). This refinement of the domain concepts allows for
clear differentiation between targeting applications, but does not lose sight of
the diversity of farm households and farming systems in homogenously
targeted groups. Wotowiec et al. stress three points: (1) domain definition
should recognize a problem focus; (2) the concept of domain should be linked
to the farming system research and development sequence; and (3)
socio-economic considerations should be included. Three types of domains are
offered in order of their .application in the farming system research and
development sequence: research domains; recommendation domains; and
diffusion domains.


Research Domains
Research domains refer to environmental (agro-ecological and
socio-economic) ranges that are defined in terms of a particular problem to be
addressed. They set the initial context for research and development
activities, often specified in terms of bio-physical factors. Broadly defined, a
research domain is an area in which agricultural development efforts are
focused to improve production and incomes of farm households. For
pragmatic reasons, initial specification of the research domain is often
geo-politically limited to a province or other administrative unit.

Within this geo-political area, further refinement of the research domain
must be specified in terms of geo-social location, soil types and cropping
potentials that distinguish different agro-ecological zones. It should be
noted, that while geographic boundaries facilitate conceptualization, they are
not absolute. There is always a degree of interspersion among research
sub-domains. The problem focus for each sub-domain should be further
narrowed because particular bio-physical and socio-economic constraints in
each agro-ecological zone limit the range of possible solutions.
Research domains correspond to the diagnosis phase of farming systems
research and development, that is, the planning of agricultural research and
extension activities. The object of the diagnostic phase is to provide an
adequate understanding of the variability in regional farming systems in
order to facilitate the targeting of homogeneous farming systems. As more is
learned about farming systems in the region, research domains can be more
narrowly defined for purposes of targeted sondeos and on-farm diagnostic trials
leading to the identification of recommendation domains.


Journalfor Farming Systems Research-Extension






THE CONCEPTUAL BASIS FOR TARGETING FARMING SYSTEMS


Recommendation Domains
A recommendation domain is a group of farm households or farming
systems with a common specific problem for whom a tested solution may be
adapted to farm household needs and potentials. It assumes a
homogeneity of the group, both socio-economically and agro-ecologically,
with reference to the problem and to the solution. A recommendation domain
cannot be established prior to problem diagnosis and solution
development. Recommendation domains are both tentative and variable in
nature, depending on problem definition and the evolution in the adaptive
research process that defines the solution. They can be modified over time
as new information is gathered and improved understanding of the problems
achieved. This usually evolves as researchers and farmers develop shared
perceptions of problems and their potential solutions. Since a recommendation
domain is limited to a specific problem and its solution, a farm household
may be involved in several recommendation domains, depending on the
number of problems and solutions addressed.
There is no recommendation domain if no solution is proposed for a
specific problem. The specification of recommendation domains cannot occur
until: 1) a specific farming system has been identified; and, 2) a solution
that fits within the particular constraints of that farming system has been found.
Often early diagnostic work (surveys, sondeos, and on-farm trials) in the
research domain will only narrow the range of the research domain in which
more focused diagnostic investigation with farmers can be pursued.

Diffusion Domains
A diffusion domain is a network of interpersonal communication that
circulates information and knowledge about solutions to problems faced by
farm households and their farming systems. A diffusion domain is not
equivalent to a recommendation domain. Sharing similar constraints does
not imply sharing of interpersonal relations. Although a diffusion domain is
logically most active at the end of the farming systems research and
development sequence, early identification of such networks can facilitate
targeting of these information flows. Who -- men, women or children -- is
involved in which communication network is important.


FARMING SYSTEM TYPOLOGIES

The use of typologies is gaining credence among farming system practioners.
Increasing recognition of important differences between farm operations
within the same area (Garrett, 1984; Wotowiec et al., 1988; Bagchee, 1990)
has led practitioners to develop methods reducing the heterogeneity of farming
systems to relatively homogeneous types (Capillon, 1986; Low, 1986a;


VoL 5, No. 2, 1995






MOORE


Marz, 1990). Farming system typologies have been most frequently developed
on a disciplinary root incorporating an interdisciplinary perspective. Early
typological frameworks building on Ruthenberg (1980) were essentially
agronomic, closely linked to the technical aspects of agricultural technologies.
These typologies tend to remain at the cropping system level with
socio-economic or animal husbandry characteristics as exogenous, when
considered. More recent typologies, such as those of Capillon (1986), Low
(1986a), and Marz (1990) have had a more economic focus. However, most
typologies developed within the farming systems tradition exhibit little
conceptual development beyond the original disciplinary considerations.
Typically, when practitioners incorporate a socio-economic element they
simply take a single indicator (usually a size referent) and designate two or
three groups (Lopez-Pereira et al., 1990; Kabay and Zepeda, 1991;
Langworthy, 1991; Franzel, 1992). Although large and small provide a
basis for obvious differences between farms, no systemic logic determines the
break point between groups. Issues of risk (Cancian, 1980), technical
efficiency (Azzam et al., 1993), the quality of resources (Park, 1993), and
the interdependence of activities conducted (Norman and Gilbert, 1982) must
be taken into account.

Assessing the farming systems research and development approach,
Bagchee (1990:129) noted that "there is very little guidance in the literature on
what exactly one can take as the factors or dimensions that differentiate
one farming system from another". Although typologies can be derived
theoretically or empirically, their use in farming systems research and
development has evolved because of initially poor pre-conceptions of existing
farming systems. Consequently, farming system typology construction has
been inductive, involving a progressive refinement of concepts for the
targeting of research and development activities. A fully transdisciplinary
theory of farming systems has yet to be developed.

The Concept of Typology
A typology is a multi-dimensional classification based on relations of
contiguity or similarity (Bailey, 1973). Like a system, it is an artificially
constructed conceptual tool facilitating comprehension and analysis. As such,
it has similar limitations. However, these tools have different, although
interrelated, purposes. Systems provide a way of thinking about the empirical
world. Typologies aid in applying those modes of thinking to the diversity of
observable conditions.
The objective of farming system typologies is to provide a categorization
and description of farming systems to facilitate recommendation domain
development. To achieve this task, a typology must identify homogeneous
farming system conditions. The characteristics described for each type of


Journal for Farming Systems Research-Extension






THE CONCEPTUAL BASIS FOR TARGETING FARMING SYSTEMS


farming system enable precise specification of more narrowly-defined research
domains and help focus the dialog between development agents and farmers
concerning problem diagnosis and solution development.
Several questions are posed in moving from the systems theoretic
framework to the identification of empirical farming systems: (1) How many
types are there? (2) What are the differentiating criteria? (3) How many criteria
are necessary? (4) How will farm households be assigned to different groups?
These questions cannot be answered at this stage. The answers are ultimately
a product of the dialog with existing conditions. However, previous farming
systems research and literature on typology construction involving farm
households can be combined to provide sufficient guidance to elaborate a
conceptual checklist to initiate that dialog.
Although the following set of criteria that form the basis for conceptual
development of a farming systems typology constitute tentative working
hypotheses, they are not formulated for explicit testing. Rather, they suggest
data collection and analyses needs to ensure the development of a useful
typology. Farming system typologies provide a context for hypothesis testing
(i.e., clearly specified research domains); working hypotheses simply guide the
process of typology development (research domain specification).
Typology development requires specifying: (1) the scope of the phenomena
to be classified; (2) the criteria used in classification; and (3) a method to apply
the criteria identifying different types. The scope of the phenomena has been
defined above. Farming systems by their nature are composed of complex
interdependencies centering on the farm household. The range of potential
diversity in farming systems requires that classifying criteria focus on
multiple dimensions of the components involved in those interdependencies.
Although a single criterion may serve to distinguish different farming systems
in a specific place at a specific time, it is not sufficient to discern the
complex interdependencies that define those different farming systems. Some
means of multiple classification should be utilized (e.g. Moore et al., 1993).
At the outset, however, a listing of potentially relevant criteria should be
specified.
Before developing the list of conceptual criteria, it should be noted that the
most common criteria used to distinguish between farming systems has been
that of environmental or agro-ecological zones (Winkelmann and Moscardi,
1982; Hart, 1982; Wotowiec et al., 1988; Jamtgaard, 1989; Gillard-Byers and
Blackie, 1990; Bagchee, 1990). At the heart of most bio-physical typologies is
a conflation of zone and system concepts. These typologies assume that
there is an equivalence between zone and system; and that bio-physical
factors of the environment reflect homogeneity in socio-economic factors at
the farm household level. Agro-ecological zones have a place in the
hierarchy of agricultural systems (see above), however, they are not a substitute
for the farm household level. In fact, Low (1986a) found that there was


VoL 5, No. 2, 1995






MOORE


greater variation between types within homogeneous agro-ecological regions
than between regions. In the approach suggested here, agro-ecological zones
provide an environmental context in which farming systems at the farm
household level are defined.

Typology Construction Criteria
The following consideration of criteria for differentiating farming systems
at the farm household level introduces social scientific literature outside of the
bio-physically-driven farming systems tradition. Although a vast literature
exists on the processes of agricultural development and differentiation of
peasant households (Garrett, 1984), this review is limited to recent literature
focusing on farming systems and farm household typologies. Realizing the
non-observability and historical nature of causal mechanisms differentiating
farming systems (Whatmore et al., 1987a; White, 1989), the focus is on farm
household and related sub-system indicators frequently linked to those
processes of development and differentiation.

Labor is the most frequently mentioned type of criteria. Social scientists
have focused on this key component of peasant farming since Lenin (1960)
and Chayanov (1966) described the dynamics of agricultural development
and differentiation early in this century. Indeed, Chayanov called these
farm households "family labor farms". The single most identified
distinguishing factor influencing farm household behavior has been
off-farm wage labor activity (Moore, 1984; Low, 1986b; Shand, 1986;
Whatmore et al., 1987b; Eboli and Turri, 1988; Tully, 1990). Consequently,
the analysis of farm household member opportunity costs is gaining wide
acceptance (Low, 1986b; Tully, 1990). In addition, labor criteria also include:
(1) a specification of the gender and age division of labor; (2) the types of
tasks assigned to these different categories; (3) the duration and seasonality of
labor allocations; (4) the intensity of the labor process; and (5) the amount and
timing of non-household labor hired or exchanged (see also: Coughenour and
Swanson, 1983; Pascon, 1986; Moock, 1986; Poats et al., 1988; Singh, 1988;
Hebinck, 1990; Marz, 1990).

Wealth often is suggested as an important indicator of differentiation. The
level of productive and consumptive durable goods indicates the cumulation of
advantages and disadvantages which the dynamics of the agricultural system
elicits over time within farm households (White, 1989). Capital now
predominates over land ownership as the key dimension of rural social
stratification. The advantage of these factors as indicators is that they
provide a measure of the potential for capitalization from non-farm (i.e., from
wage labor and non-agricultural activities) as well as farm sources. These
criteria include: (1) the possession of farm equipment; (2) debt or credit


Journal for Farming Systems Research-Extension






THE CONCEPTUAL BASIS FOR TARGETING FARMING SYSTEMS


levels; (3) non-agricultural investments in businesses or housing; and (4) the
possession of consumption durables (see also: DeWalt and DeWalt, 1980;
Garrett, 1984; Low, 1986b; 1989; Hebinck, 1990; Tully, 1990).

Land has been a standard criterion for distinguishing between farm
households. As the fundamental environmental resource involved in all
agricultural systems, it is often included in the design of typologies with such
designations as small, medium and large. Due to this overly simplistic
usage, it is coming into increasing disrepute (Whatmore et al., 1987a; White,
1989; Bagchee, 1990). Size categories on their own do not account for land
quality in terms of soil type, depth, water availability or slope. Tenure status
is also an important dimension indicating the social relations governing the
use and output of accessible land. However, different forms of tenure status,
in themselves, do not necessarily indicate the value of the land to each of the
partners in the relationship (Whatmore et al., 1987a). In addition,
intra-household relations concerning the ownership and use of land should
also be taken into account (Moock, 1986). Criteria indicating the land
resources available to each household include: (1) the amount of land
operated; (2) the quality of that land; and (3) the social relations governing
the use of the land (see also: Garrett, 1984; Pascon, 1986; Hebinck, 1990).

Household demographic composition has gained increasing attention as a
key criterion distinguishing farm households and their modes of functioning.
It has long been recognized that stage in the family life cycle and extended
family structures have an important impact on how the farm is run
(Chayanov, 1966). The number, age and gender of household members has a
direct impact on both labor resources and consumption needs (Pascon,
1986). Since the farm household is the primary unit of analysis, it is
important to specify what constitutes a household. Household studies
continue to point out not only that family may not be an adequate concept, but
that household is also an ambiguous concept (Netting et al., 1984; Moock,
1986; White, 1989). Here we define the household as composed of all
persons residing in the producing and consuming unit. This may include
non-family-related persons. Non-resident family members also may be
significant contributors/consumers. These need to be accounted for
separately. Demographic criteria should include: (1) the number, age and
gender of all household and non-resident family members; and (2) their
family (or non-family) relationships (see also: Crummett, 1987; Eboli and
Turri, 1988; Poats et al., 1988; Wilk, 1989).

Crop composition and production techniques are the most frequent criteria
used to differentiate farming systems. Land use patterns, whether perennial,
annual or seasonal provide important information on how a farming system is


VoL 5, No. 2, 1995






MOORE


managed (Ruthenberg, 1980). Types of crops grown further indicate the
objectives -of production whether for home consumption, animal feed or sale.
Crop production techniques and the use of purchased inputs are often
distinctive of different farming systems. Crop yields, although essentially a
performance indicator, are sometimes used to differentiate types. These criteria
include: (1) the area committed to each crop produced; (2) the rotation
patterns of these crops; (3) land preparation techniques; (4) the use of
purchased inputs; (5) cultivation practices; (6) harvesting techniques; (7) crop
yields; and (8) the end use of each crop (see also: Hart, 1982).

Livestock composition and production techniques are less often mentioned,
but equally important criteria (McDowell and Hildebrand, 1980). Herd size
and stocking rate indicate the scale and intensity of livestock activities. Animal
purchases and sales allow for distinguishing between market and
consumption orientations. The combination of types of animals may
suggest multiple purposes. The distinction between local and improved
breeds, the use of purchased feeds, and the use of veterinary services
indicate production techniques. An important additional criteria often
missed is the extent and form of herd ownership. Livestock criteria should
include: (1) the number and type of livestock; (2) stocking rate; (3)
purchases and sales of livestock; (4) the feeding schedule: (5) the use of
purchased feeds and veterinary services; and (6) the social relations concerning
the ownership and management of livestock.

Farm and family goals are frequently mentioned as key determinants of
farming system dynamics. These goals provide the ultimate motivation for
engaging in agricultural production. Although manifested individually,
they involve culturally-defined values of the community, as well as of
particular productive activities (Long, 1984). The real potential for choice in
these matters may be mitigated by the need for physical survival within
culturally-defined norms. Like other causal mechanisms, such values are
difficult to observe directly. Although the frequently hypothesized dichotomy
between self-sufficient peasants and market-oriented producers is rarely
'found, the uses of produced outputs often serve as valuable indicators of
household goals. Criteria to indicate these goals include: (1) the types of
production pursued; (2) the level and form of self-sufficiency; and (3)
expectations for children (see also: Capillon, 1986; Eboli and Turri, 1988;
Long and van der Ploeg, n.d.).
Closely linked with farm and family goals is the historical trajectory of
the farm operation. Although seldom used in typology development because
of the retrospective data collection and complex analyses involved, it is
important to place the farming system on an evolutionary path to map the
dynamics of system development (Capillon, 1986: Diaz et al., 1990). Criteria


Journal for Farming Systems Research-Extension







THE CONCEPTUAL BASIS FOR TARGETING FARMING SYSTEMS


for indicating such trajectories could include all criteria mentioned above
for previous developmental stages of the farm household systems.


SUMMARY AND CONCLUSIONS

The concepts and criteria discussed in this paper provide tools for the
targeting and implementation of agricultural research and development.
Two levels of conceptual tools are involved. The first provides a way of
thinking about the dynamic totality of agricultural development. The second
channels this way of thinking toward the applied targeting of specific farming
systems. Both aid in the diagnosis of farmer problems and how those
problems may be resolved.
The concept of system is central to applied agricultural research and
development. Farming operations must be conceptualized and examined
as systemic wholes involving farm household members allocating their time
and resources to three interdependent processes (crops, livestock, and
off-farm activities) in order to attain their culturally and materially defined
goals. These processes have been defined as sub-systems nested in and
composing, with household resources, the farming system.
Furthermore, farming systems exist within both bio-physical and
socio-economic environments. Systematically interlinked, these
environments compose a hierarchy of agricultural systems. Within this
hierarchy, the ultimate decision to adopt, adapt or reject technological
innovations is made at the farm household level. Nevertheless, important
characteristics of the agricultural system as a whole need to be taken into
account; in particular, the agro-ecological and socio-economic conditions
in which farming systems exist.
This farming systems conceptual framework provides a way of thinking
about the systematic interdependence of both the internal and external
components of farming operations. However, considerable diversity exists in
the real world of farming. Additional conceptual tools are necessary to
bridge the gap between this way of thinking and real world applications.
Research domains need to be clearly specified and ultimately narrowed to
specific farmer problems and their solutions.
Bio-physical features conditioning production potentials provide the
initial basis for research domain specification. Agro-ecological zonage of
geo-politically defined research domains should be developed to account for
these sources of differentiation. However, not all farm households are
equally endowed with resources; nor do they combine their resources in the
same manner. The appropriate targeting of farming systems for
technological innovations also requires the identification of these differences
and the limits which they impose.


VoL 5, No. 2, 1995






MOORE


Within the context of agro-ecologically defined research domains, the
interdependence of resources and productive activities at the level of the farm
household provide the systemic logics distinguishing one type of farming
system from another. The potential for a technology to be a solution to a
farmer's problem depends on: its applicability to the specific
agro-ecological conditions in which the farming system exists; and its
compatibility with the socio-economic capacities and objectives of the
individual farm household managing that farming system. Narrowing research
domains to a farming systems typology is the final span in the bridge from
conceptual understanding of farming systems to applied problem diagnosis and
solution development.
The specification of a typology of farming systems does not complete
the targeting cycle. The bridge must be crossed. The role of the typology is to
focus investigation and dialog with farmers within clearly-defined and
homogeneous research domains for problem diagnosis, testing and solution
development. Only then, can targeted recommendation domains be established
and messages circulated through diffusion domain networks.


REFERENCES


Azzam, A.M., S.M. Azzam and M. Moussaoui. 1993. Categorizing Farms as Large or Small: An
Econometric Application to Chaouia Farms in Morocco. pp. 14. Programme Aridoculture.
Centre R6gional de la Recherche Agronomique, Institut National de la Recherche Agranomique.
Settat, Morocco.
Bagchee, Aruna. 1990. Agricultural Development in India: An Assessment of the FSRE Approach.
Journal for Farming Systems Research-Extension, 1(2):123- 136.
Bailey, Kenneth D. 1973. Monthethic and Polythetic Typologies and Their Relation to
Conceptualization, Measurement and Scaling. American Sociological Review
38(February):18-33.
Baker, D.C. and D.W. Norman. 1990. The Farming Systems Research and Extension Approach to
Small Farmer Development. In M.A. Altieri and S.B. Hecht (eds.) Agroecology and Small
Farm Development. CRC Press. Boca Raton, Florida
Brossier, Jacques. 1987. Systane et system de production: note sur ces concepts. Cahiers des
Sciences Humaines, 23(3/4):377-390.
Brush, Stephen B. and B.L Turner II. 1987. The Nature of Farming Systems and Views of Their
Change. In Turner I, B.L. and S.B. Brush (eds.) Comparative Farming Systems. The
Guilford Press. New York.
Cancian, Frank. 1980. Risk and Uncertainty in Agricultural Decision Making. In P. Barlett (ed.)
Agricultural Decision Making. Academic Press. New York.
Capillon, Alain. 1986. A Classification of Farming Systems, Preliminary to an Extension Program: A
Methodology. pp. 17. Farming Systems Research Paper Series No. 11. Kansas State University.
Manhattan, Kansas.
Chayanov, A 1966. Peasant Farm Organization. In D. Thomas, B. Kerblay and R. Smith (eds.) A.V.
Chayanov on the Theory of Peasant Economy. American Economic Association. Homewood,
Illinois.
Conway, Gordon R. 1986. Agroecosystem Analysis for Research and Development. Winrock
International. Bangkok, Thailand.


Journal for Farming Systems Research-Extension








THE CONCEPTUAL BASIS FOR TARGETING FARMING SYSTEMS


Coughenour, C. Miklon and Louis Swanson. 1983. Work Statuses and Occupations of Men and
Women in Farm Families and the Structure of Farms. RuralSociology, 48(1)23-43.
Crummett, Maria de los Angeles. 1987. Class, Household Structure, and the Peasantry: An
Empirical Approadl. The Journal of Peasant Studies, 14(3)363-379.
DeWalt. Billie R. and Kathleen Musante DeWalt. 1980. Stratification and Decision Making in the
Use of New Agricultural Technology. In P. Barlett (ed.) Agricultural Decision Making:
Anthropological Contributions to Rural Development. Academic Press. Orlando, Florida.
Diaz, M., 0 Sotomayor and J. Berdegui. 1990. Sustainability and Productivity of Mapuche Farming
Systems Journal }f Farming Systems Research-Extension, 1(1)109-123.
Dowling, Grahame R. 19X7. Dialogue on Systems as Clusters: Identifying Systems. Behavioral
Science, 32:149-152.
Eboli, Maria G. and Enrico Turri. 1988. Toward a Behavioral Model of Multiple-Job-Holding Farm
Families. Agricultural Economics, 2:247-258.
Food And Agriculture Organization. 1989. Farming Systems Development: Concepts,
Methods, Applications. Rome, Italy.
Franzel, Stephen. 1992. Extension Agent Surveys for Defining Recommendation Domains: A Case
Study from Kenya. Journal for Farming Systems Research- Extension, 3(1):71-86.
Friedland, W.F., A.E. Barton and R.J. Thomas. 1981. Manufacturing Green Gold Capital Labor
and Technology in the Lettuce Industry. Cambridge University Press. New York, New
York.
Garrett, Patricia. 1984. The Relevance of Structural Variables for Farming Systems Research.
Rural Sociology, 49(4):580-589.
.Gibbs, Christopher G.N. 1985. Agricultural Systems Research in Asia: A Comparative
Discussion of Human Ecology, Agroecosystems Research, Farming Systems Research, and
Cropping Systems Research. Agricultural Systems, p. 77-91.
Gillard-Byers, Thomas E. and Malcolm Blackie. 1990. Zoning Survey: Improving the Efficiency of
Farming Systems Research-Extension Diagnostic and Field Acivities Journal for Farming
Systems Research-Extension, 1(2):105-121.
Hart, Robert D. 1982. An Ecological Systems Conceptual Framework for Agricultural Research
and Development. In Shaner, W.W., P.F. Philipp and W.R. Schmehl (eds) Readings in Farming
Systems Research and Development. Westview Press. Boulder, Colorado.
Hebinck, Paul. 1990. The Agrarian Structure of Kenya: State, Farmers and Commodity
Relations. Verlag Breitenbach Publishers. Fort Lauderdale, Florida.
International Center for Agricultural Research in the Dry Areas. 1989. Sustainable Agriculture
for the Dry Lands: ICARDA's Strategy and Medium- Term Plan. ICARDA. Aleppo, Syria.
Jamtgaard, Keith A 1989. Targeing Production Systems in the Small Ruminant CRSP: A Typology
Using Cluster Analysis. In Constance M McCorkle (ed.) The Social Sciences in International
Agricultural Research: Lessons from the CRSPs. Lynne Rimner Publishers. Boulder,
Colorado.
Kabay, M.B. and Lydia Zepeda. 1991. Crop-Enterprise Selection of Farming Systems in Eastem
Gambia. Journal for Farming Systems Research- Extension, 2(1):87-96.
Langworthy, Mark. 1991. Measurements of Economic Viability in Cape Verde. Journal for
Farming Systems Research-Extension, 2(1):109-124.
Lenin, V. 1960. The Development of Capitalism in Russia. Volume 3 of Collected Works.
Lawrence and Wishart. London.
Long, Norman. 1984. Introduction. In N. Long (ed.) Family and Work in Rural Societies:
Perspectives on Non-Wage Labour. Tavistock Publications. New York.
Long, Norman, and Jan Douwe van der Ploeg. n.d. Heterogeneity, Actor and Structure: Towards a
Reconstitution of the Concept of Structure. pp. 29. Wageningen University. Wageningen, The
Netherlands.
L6pez-Pereira, M.A, T.G. Baker, J.H. Sanders and D.H. Meckenstock. 1990. Farming Systems
and Adoption of New Agricultural Technologies: An Economic Evaluation of New Sorghum
Cultivars in Southem Honduras. Journal for Farming Systems Research-Extension,
1(2):81-103.


Vol 5, No. 2, 1995







36 MOORE

Low, Allan. 1986a. On-Farm Research and Household Economics. In J.L. Moock (ed.)
Understanding Africa's Rural Households and Farming Systems. Westview Press.
Boulder, Colorado.
Low, Allan. 1986b. Agricultural Development in Southern Africa: Farm-Household Economics and
the Food Crisis. James Currey. London.
Marcotte, Paul and Louis E. Swanson. 1987. The Disarticulation of Farming Systems Research
with National Agricultural Systems: Bringing FSR Back In. Agricultural Administration and
Extension, 27:75-91.
Marz, Ulrich. 1990. Farm Classification and Impact Analysis of Mixed Farming Systems in Northern
Syria. Wissenshaftsverlag Vauk Kiel. Kiel, Germany.
McDowell, R.E. and P.E. Hildebrand. 1980. Integrated Crop and Animal Production: Making
the Most of Resources Available to Small Farms in Developing Countries. The Rockefeller
Foundation. New York.
Moock, Joyce Lewinger. 1986. Understanding Africa's Rural Households and Farming
Systems. Westview Press. Boulder, Colorado.
Moore, Keith M. 1984. The Household Labor Allocation of Farm-Based Families in Wisconsin.
University Microfilms International. Ann Arbor, Michigan.
Moore, K.M., F. Nassif, A. Sefrioui, and R. Riddle. 1993. Aridoculture Baseline Study and
Farming Systems Typology. Serie B 05/1. Programme Aridocilture. Centre Regional de la
Recherche Agronomique. Intitut National de la Recherche Agronomique. Settat, Morocco.
Netting R.M., R.R. Wilk and E.J. Amould. 1984. Introdudion. In R.M. Netting, R.R. Wilk and
E.J. Amould (eds.) Households: Comparative and Historical Studies of the Domestic Group.
University of California Press. Berkeley, California.
Norman, D.W. and Elon Gilbert. 1982. A General Overview of Farming Systems Research. In
Shaner, W.W., P.F. Philipp and W.R. Schmehl (eds.) Readings in Farming Systems Research
and Development. Westview Press. Boulder, Colorado.
Oasa, Edmund K. and Louis E. Swanson. 1986. The Limits of Farming Systems Research and
Development: Should Development Administrators Be Interested? Agricultural Administration,
23:201-221.
Park, Thomas K. 1993. Risk and Tenure in Arid Lands: The Political Ecology of Development in the
Senegal River Basin. The University of Arizna Press. Tucson.
Pascon, Paul. 1986. 30 Ans de Sociologie du Maroc. Bulletin Economique et Social du Maroc.
Sochepress. Casablanca, Morocco.
Poats, S.V., M. Schmink and A. Spring. 1988. Gender Issues in Farming Systems Research and
Extension. Westview Press. Boulder, Colorado.
Ruthenberg, Hans. 1980. Farming Systems in the Tropics. Third Edition. Clarendon Press.
Oxford, Great Britain.
Shand, R.T. 1986. Conclusions. In R.T. Shand (ed.) Off-Farm Employment in the Development of
Rural Asia, Volume One. National Centre for Development Studies. Australian National
University.
Shaner, W.W. 1982. Introduction. In Shaner, W.W., P.F. Philipp and W.R. Schmehl (eds.)
Readings in Farming Systems Research and Development. Westview Press. Boulder,
Colorado.
Shaner, W.W., P.F. Philipp and W.R. Schmehl. 1982. Farming Systems Research and Development:
Guidelines for Developing Countries. Westview Press. Boulder, Colorado.
Simmonds, Norman W. 1985. Farming Systems Research: A Review. World Bank Technical
Paper No. 43. International Bank for Reconstruction and Development. Washington, D.C.
Singh, Ram D. 1988. Economics of the Family and Farming Systems in Sub-Saharan Africa:
Development Perspeives. Westview Press. Boulder, Colorado.
Tully, Dennis. 1990. Labor and the Adoption and Impact of Farm Tedmology. In D. Tully (ed.)
Labor and Rainfed Agriculture in West Asia and North Africa. Kluwer Academic
Publishers. Dortrecht, The Netherlands.
Whatmore, S., R. Munton, J. Little and T. Marsden. 1987a. Towards a Typology of Farm
Businesses in Contemporary British Agriculture. Sociologia Ruralis, 27(1):21-37.


Journal for Farming Systems Research-Extension







THE CONCEPTUAL BASIS FOR TARGETING FARMING SYSTEMS


Whatmore, S., R. Munton, T. Marsden and J. Little. 1987b. Interpreting a Relational Typology of
Farm Businesses in Southern England. Sociologia Ruralis, 27(2/3):103-122.
White, Benjamin. 1989. Problems in the Empirical Analysis of Agrarian Differentiation. In G.
Hart, A. Turton, and B. White (eds.) Agrarian Transformations: Local Processes and the
State in Southeast Asia. University of California Press. Berkeley, California.
Whyte, William Foote. 1983. Systems Thinking for Understanding the Small Farm Enterprise. In
Whyte, W.F. andD. Boyntcn (eds.) Higher Yielding Human Systemsfor Agriculture. Cornell
University Press. Ithaca, New York.
Wilk, Richard R. 1989. The Household Economy: Reconsidering the Domestic Mode of Production.
Westview Press. Boulder, Colorado.
Winkelmann, Donald and Edgardo Moscardi. 1982. Aiming Agricultural Research at the Needs of
Farmers. In Shaner, W.W., P.F. Philipp and W.R. Schmehl (eds.) Readings in Farming
Systems Research and Development. Westview Press. Boulder, Colorado.
Wotowiec, Jr. P., S.V. Poats and P.E. Hildebrand. 1988. Research, Recommendaticn and
Diffusion Domains: A Farming Systems Approach to Targeting, In S.V. Poats, M. Sdhmink
and A. Spring (eds.) Gender Issues in Farming Systems Research and Extension. Westview
Press. Boulder


VoL 5, No. 2, 1995









ADOPTION OF SOYBEAN: A COMPARATIVE
ANALYSIS OF CULTURAL PRACTICES IN
ZAIRE AND NIGERIA


Dennis A. Shannon, Mwamba Kalala M., Kubengu Mudilamika and
Mpoy Mudiamvita C.1


ABSTRACT

The suitability of soybean for peasant agriculture in sub-
Saharan Africa has been in dispute for many years. This
controversy has affected the allocation of resources for
research in soybean. In order to justify research in soybean, a
survey was conducted to document soybean production and
utilization in Gandajika, Zaire. Soybean cultivation was
widespread in the area. Cropping practices were described
and compared with cropping practices in areas of Nigeria
where soybean is grown. The cropping practices in the two
countries reflected similarities and differences in the physical
environment, major crops and economic conditions.
Similarly, farmers' perceptions and perceived constraints
relating to soybean production reflected differences in the
physical and economic environments. Soybean culture was
integrated into traditional farming systems in both Zaire and
Nigeria. These observations suggest that soybean was
relevant to farmers in these regions, and that research to
improve soybean production was needed. This survey served
to justify continued research in soybean, thus demonstrating
the utility of multi-disciplinary crop surveys for orienting
research programs to the needs of farmers.


Assistant Professor, Department of Agronomy and Soils, 202 Funchess Hall,
Auburn University, Alabama 36849-5412, U.S.A., formerly Agronomist,
International Institute of Tropical Agriculture, Ibadan, Nigeria; formerly, Institut
National pour 1'Etude et la Recherche Agronomique, B.P. 2037, Kinshasa I,
Zaire; and Research Assistant and Former Research Assistant, respectively,
Programme National L6gumineuses, B.P. 22, Mbuji Mayi, Kasai Oriental, Zaire.


VoLS, No.2, 1995






D.A. SHANNON, K.M. MWAMBA, M. KUBENGU, M.C.MPOY


INTRODUCTION


Until recently, many agriculturalists familiar with Africa believed that soybean
(Glycine max) was not suited to peasant agriculture in sub-Saharan Africa.
This belief was predicated on the assumptions (1) that African farmers would
not eat soybean (Vanneste, 1986); and (2) that agronomic limitations rendered
the crop unsuited to low-resource farmers in the lowland humid and sub-humid
tropics. Unlike most other legume crops, soybean varieties developed in
temperate regions were not able to nodulate and fix nitrogen with Rhizobium
bacteria indigenous to African soils (Dashiell et al., 1987). Secondly, when
seed of these varieties were stored under conditions of high humidity and
temperature, they rapidly lost their viability. These two constraints severely
restricted the crops' utility to small-scale farmers who could not obtain
inoculant and proper seed storage facilities. Whingwiri (1987) reported that
soybean production was "not very attractive" to peasant farmers in Zimbabwe
owing to a number of constraints including poor nodulation, lack of knowledge
about inoculants, shattering and low yields. However, soybean varieties are
now available that combine high yield potential with improved seed longevity,
a capacity to nodulate with a wide range of rhizobium strains common to
tropical African soils and resistance to shattering and common diseases
(Dashiell et al., 1987; IITA, 1992). This has greatly improved the chances of
soybean being accepted under a wide range of tropical conditions.
The assumption that farmers will not eat soybean has been based upon
attitudes and beliefs regarding soybean foods. Common objections to soybean
related to cooking time, taste, unfamiliarity with acceptable uses, and belief by
some that the crop was poisonous (Weingartner et al., 1987). Others felt that
due to perceived problems limiting home consumption, as well as technical
constraints in growing the crop, soybean was mainly suited to large-scale
commercial production destined for industrial processing. Doubts concerning
the potential adoption of soybean were so prevalent that by 1986, the
International Institute of Tropical Agriculture (IITA), where much innovative
research had been undertaken to address problems affecting soybean in the
tropics, relegated soybean to a secondary position among mandate crops (p. 44
in IITA, 1987). Two years later, the Institute ranked the projected importance
of soybean and the productivity of research in soybean as low (IITA, 1988).
Although attitudes toward soybean may be changing, misgivings persist
regarding the crop's appropriateness to small-scale farming in sub-Saharan
Africa.
Sub-Saharan Africa is vast, with a range of agro-ecological
environments and a multitude of ethnic cultures with their own traditions,


Journal For Farming Systems Research-Extension






ADOPTION OF SOYBEAN ZAIRE AND NIGERIA


foods, crops, cropping systems and economic conditions. It is therefore risky to
generalize from experiences at one location to other locations within the region.
Yet, it is necessary to anticipate the likelihood of adoption of technological
interventions in order to make rational choices in the prioritization of research
activities. For research activities to be cost-effective, the technologies
developed must be adopted by many people over large geographic areas. Thus,
evidence of the adoption of soybean as a crop at a single location, although
positive, is not adequate to predict adoption in other areas, or to justify a major
research program. However, evidence of adoption in areas differing in ethnic
cultures, foods, crops, cropping systems and economic environment may be
seen as strong evidence that the crop is likely to be adopted in other areas
within the region where it is introduced.
Shannon and Mwamba (1994) presented evidence that soybean has
been integrated into the diet of farmers in Zaire, Nigeria and other sub-Saharan
countries. Commonly cited problems with soybean preparations were based
upon preconceived notions of how the crop would be utilized. Those who
predicted that the crop would not be adopted failed to anticipate the
development of indigenous preparation methods to which these problems did
not apply. In Gandajika, Zaire, soybean flour is used in traditional maize
porridge, in the maize-cassava staple, nshima, or added to vegetables. The
grain is also roasted and eaten whole. In Nigeria, soybean is used in a number
of traditional foods, the most important being as a substitute for locust bean
(Parkia spp.) in daddawa, for melon seed (Citrullus lanatus) in egussi stew,
and as a supplement to cowpea (Vigna unguiculata) in moin moin and akara
(Pfeiffer, forthcoming; Smith et al., 1993). In Gandajika, soybean ranked
second to cowpea in terms of area planted to grain legumes (Shannon and
Mwamba, 1994). Virtually all production is used for human food, with one
third being consumed directly within producing households. In the Tiv-
speaking area of Benue State, Nigeria, the average farmer interviewed produced
over 400 kg in 1989 (Smith et al., 1993), a substantial production for small-
scale farmers.
Documentation on existing cropping systems involving soybean in
sub-Saharan Africa is rare, and mostly limited to socio-economic rather than
agronomic information. However, the existence of established cropping
practices adapted to specific agro-ecological environments and farming systems
is evidence of adoption of the crop. In this paper, examples of the cultivation of
soybean by low-resource farmers in Zaire and Nigeria are discussed in relation
to the larger issue of soybean adoption in sub-Saharan Africa. Established
cropping practices involving soybean grown by low-resource farmers at one
location in Zaire are described and compared to cropping practices reported
from Nigeria. Similarities and differences in soybean cropping practices in the
two countries are analyzed according to crop and environmental characteristics,


Vol 5, No.2, 1995






D.A. SHANNON, K.M. MWAMBA, M. KUBENGU, M.C.MPOY


existing cropping systems, economic considerations and farmers perceptions in
the two countries. The usefulness of simple crop surveys is also discussed.


GANDAJIKA SURVEY

The history of research and extension on soybean in the Kasai Regions of south
central Zaire was summarized by Shannon and Mwamba (1994). Soybean
adoption occurred largely as a result of efforts in Western Kasai which began
about 1960 and included active promotion of soybean cultivation, research into
appropriate ways to prepare soybean foods, village-level education in
preparation of soybean foods, and emphasis by medical institutions on using
soybean to combat protein deficiency in children. Mills were established by
foreign missionaries, where the grain was purchased, roasted and ground into
flour for resale. Seed of adapted soybean varieties were supplied by the
experiment station at Gandajika in Eastern Kasai. In 1985, a national grain
legume research program (Programme National Ligumineuses) was established
with headquarters at Gandajika. Questions were raised about how much
research effort should be allocated to soybean relative to other grain legume
crops. At issue was whether priority should be given to widely-grown
traditional crops such as cowpea, with major production problems (insect pests)
for which low-cost solutions were not readily available, or to the relatively new
crop, soybean, which had not yet been widely adopted, but which was being
widely promoted within the country. Rapid gains in the productive potential of
soybean were possible through the introduction of improved varieties developed
in a similar ecology in Nigeria by IITA. In 1986, an external evaluation team
questioned the emphasis placed by the national program on soybean, citing
both the relatively small area in soybean production compared to more
established grain legume crops, and doubts concerning the potential adoption of
soybean as a food. Only limited anecdotal information was available on
soybean production in Zaire, and the extent of adoption was unknown.
Documentation of soybean adoption was needed to justify a major research
effort in soybean. In Gandajika, soybean was seen planted during the 1985/86
and 1986/87 growing seasons. A survey was conducted to document the
importance of soybean in the area and to learn about soybean production
practices.
Gandajika may be considered an area of secondary adoption, since it
had benefited mostly indirectly from the extension efforts in Western Kasai.
The services available in Western Kasai for milling the grain were not
accessible to Gandajika farmers, who had to rely on their own resources to
process and market the crop. Thus Gandajika was a good area for assessing the
potential for adoption of soybean, because external incentives were limited.


Journal For Farming Systems Research-Extension






ADOPTION OF SOYBEAN ZAIRE AND NIGERIA


Gandajika is a zone (equivalent to a county in North America) in the
Eastern Kasai Region of Zaire. It is located in Guinea savanna (Vanhamme et
al., 1955), with a mean annual rainfall of approximately 1450 mm distributed
between late August and mid-May. Annual crops are grown in two seasons per
year, September to January being referred to as Season A, and February to May
as Season B. Maize (Zea mays) and cassava (Manihot esculenta) are the
principal crops grown. Soils are mildly acid, predominantly Alfisols, with
texture ranging from sand to sandy clay loam.


Methods
A total of .115. farmers in 9 groupements or village groups within the
Gandajika Zone were interviewed in April 1987 to obtain information on the
production and utilization of soybean and farmer knowledge of, and attitudes
toward, soybean. The survey was conducted in villages along several roads
radiating from the city of Gandajika. The population consisted of three tribal
groups, the Kaniok, Luba and Nsonge, and reflect variations in the cropping
practices observed in the area. The survey was carried out by technicians
trained in agronomy and rural sociology at the secondary and university level.
A questionnaire was drawn up in French, but the interviews were open-ended
and conducted in the Tshiluba language. Because reliable census data were not
available, interviewers circulated in the fields and interviewed farmers as they
were encountered. Although the initial intent was to obtain equal numbers of
farmers who had or had not grown soybean, it was discovered that all the
farmers interviewed had grown the crop.
Plant spacings were measured by the interviewer. Production was
reported by farmers in units of meca (approximately 3 kg) and this, together
with information on the area planted, was used to estimate yields. Additional
details of the survey methods were presented in Shannon and Mwamba (1994).
Observations and farmer responses regarding growing of the soybean
crop by small-scale farmers are presented in this paper. Information regarding
its importance and use were presented by Shannon and Mwamba (1994).


Results and Discussion
Almost none of the farmers interviewed could name the variety they
planted, but most grew a variety requiring three months from planting to
harvest. The growing period and field observation suggested that the majority
planted SJ 127, seed of which was produced at the experiment station at
Gandajika. This variety is short-statured, has yellow seed, abundant gray
pubescence and matures in 85-95 days. Jupiter, mentioned by name by three
individuals, has green seed at Gandajika, tawny pubescence and matures in
120-125 days.


Vol 5, No.2, 1995






D.A. SHANNON, K.M. MWAMBA, M. KUBENGU, M.C.MPOY


Soybean was usually cultivated by the men (64.4 %) or by the entire
family (30.4 %). Individual women (widowed or divorced) accounted for 4.3 %
of fields surveyed. Most farmers planted soybean in small plots near the
household, indicating its place as a minor crop. However, .23.5 % planted
soybean in the larger fields away from the household, which suggests that the
crop has some commercial importance to these farmers.
Agronomic practices: Soybean was most often planted sole (79.1 %
of respondents), but 15.7 % of farmers intercropped with maize. Another 5.2
% of farmers sometimes planted soybean sole and sometimes intercropped.
Soybean most often followed maize in the rotation (64.4 %). Other crops
included cowpea (6.9 %), cassava (4.3 %), cotton (3.5 %), peanut (Arachis
hypogaea, 2.6 %) or fallow (3.5 %).
Not all farmers selected the soil on which they planted soybean (Figure
1). Of the approximately one third who did select soils, their choices suggest a
knowledge that soybean did not tolerate low fertility soils as well as some other
crops, such as cassava and cowpea.
Fertilizer was scarce and was applied almost exclusively on maize.
Only 10.4 % reported having ever applied fertilizer on soybean and none had
done so in the season when interviewed.
Farmers planted soybean on flat seedbeds (55.7 %), ridges (37.4 %) or
beds (6.1 %). Flat seedbeds were chosen to permit a high plant density and to
save time (Figure 2). Ridges were chosen to facilitate weeding and to improve
fertility and moisture conditions for the crop. Beds were chosen to facilitate
intercropping with maize.
Soybean was planted in two seasons per year. Higher yields may be
expected in the Season A (September to January) because of more abundant
rainfall. However, better drying conditions at harvest ensure that better quality
seed will be harvested in Season B (January to May), provided sufficient rain is
obtained during grain filling for the crop to produce grain. When asked to
indicate which was the better season in which to grow soybean, farmers were
almost equally divided in their choice. Given the problems of seed longevity in
soybean (Dashiell et al., 1987), it is easy to understand their apparent
ambivalence regarding the two seasons.
Soybean was planted in September (65.2 %) and October (24.3 %) in
Season A and harvested in December (54.7 %) and January (32.2 %). It was
planted again in January (46.9 %) and February (51.3 %), in Season B, and
harvested in April (35.6 %) and May (61.7 %). Planting of soybean coincided,
in Season A, with the period when the farmer completed seeding and began
weeding his primary crop, maize (Figure 3). Harvest of Season A soybean took
place when the farmer was beginning his maize harvest, preparing land and
planting Season B maize and cotton (Gossypium hirsutum). Planting of Season
B soybean coincided with planting of Season B maize, but soybean harvest


Journal For Farming Systems Research-Extension








^

'o
% of 115 Farmers


Rich
19.5 %


NO
64.3 %


YES
35.7 %


Clay-Loam'
%


% of 41 Farmers


Clayey
32.0 %



Sandy 4.8 %
Sandy-Loam
4.8 %
B Previously
Black Cropped
9.7 % 9.7 %


Soil Type Chosen for Soybean


Choose Soil?







D.A. SHANNON, K.M. MWAMBA, M. KUBENGU, M.C.MPOY


Figure 2

% of Farmers
70

60

50

40
SHigh Density
30 m Easier to Weed
I Less Weeds
20 Moisture/Fertility
M Saves Time
10 9MII Intercropping
i Less Lodging
0
Seedbed: FLAT RIDGES BEDS
Farmers: 64 43 7


preceded that of maize. From the standpoint of labor use, it may have been
more convenient to plant in Season B.
Only 2.6 % of our sample saved their own seed. The remainder either
purchased grain for planting in the local market (54.0 %), or obtained seed
from a neighbor (24.3 %), the research station (13.0 %) or other family
members (6.1 %). This situation was typical of other grain legumes grown in
the region, such as cowpea and peanut. Harvested grain was either consumed
or sold to pay off debts or in response to financial emergencies. At planting
time, farmers purchased seed if they had cash. If not, no grain legumes were
planted on their farm.
Between row spacings of sole cropped soybean ranged from 20 cm to
100 cm; within row spacings ranged from 10 cm to 50 cm. The most popular
spacings were 30-40 cm between rows (54.8 %) and 20-30 cm within the row
(77.9 %). These spacings applied generally to flat seedbeds and were
consistent with recommended practices (Vanneste, 1986). Between-row
spacings of 50-80 cm (22.0 %) were generally used when planting on ridges,
while the narrow spacings were used on flat seedbeds. Seeds per hill ranged
from two to seven, with 3 (37.3 %), 4 (28.6 %) and 2 (24.2 %) seeds per hill
being most popular.
These seeding rates gave densities often exceeding 300,000 plants/ha.
Results of a plant density trial conducted on-station indicate that 200,000


Journal For Farming Systems Research-Extension










% of 115 Farmers
250,


200


150


100


50


0


Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul


LAND CLEARING
I WEEDING


TILLAGE
r- HARVEST


= PLANTING


*1






D.A. SHANNON, K.M. MWAMBA, M. KUBENGU, M.C.MPOY


plants/ha are adequate to obtain satisfactory yields ( Shannon et al., 1988). The
short stature of the popular variety, SJ 127, especially on the degraded soils
near many households, may have influenced some farmers to adopt high
densities.
With intercropping, between-row spacings ranged from 20 to 80 cm,
with 80 cm being the most frequently recorded (22.2 %) followed by 60 cm
(16.7 %) and 40 cm (16.7 %). Within-row spacing ranged from 20 cm to 100
cm, but was usually between 30 cm and 20 cm (77.7 %). Two (50.0 %) to three
(38.8 %) seeds were normally planted per hill. Densities were consequently
highly variable.
Yields: Average yields were estimated at 412 kg/ha in the previous season,
with a range in yield of 90 to 2400 kg/ha. The latter was remarkable since no
fertilizers were used.
Area Planted: Field size varied from 20 m2 to 1 ha, but averaged 0.18 ha is
Season A and 0.11 ha in Season B. Not all farmers planted soybean in Season
A and of those who did, not all were able to estimate the area they cultivated in
Season A. Consequently, it must not be inferred that farmers planted more area
in Season A than in Season B.
Farmer evaluation of soybean: Most farmers reported no problems growing
soybean (62.6 %). Insects were considered a problem by a quarter of farmers
(24.3 %). Poor germination, an unproductive variety and shattering were also
mentioned. To counter insect pests, some farmers applied ashes or insecticides
to the plants. Timely harvest was necessary for their shattering-prone variety
and rogueing was practiced to minimize the spread of the shimbu mealybug
(Crypticerya spp.).
Most farmers cited marketing as a major constraint limiting the
amount of soybean grown (82.6 %). Half stated that they had little time to
grow soybean. The problem of drying the soybean following Season A was
cited by 38.3 % of respondents. Lack of a good variety was cited by a quarter of
the farmers (27.8 %), as was over-abundant rainfall (27.0 %). Other
constraints mentioned by a minority of farmers were shattering (10.4 %),
ignorance (7.8 %), difficulty in eating soybean (6.1 %), insect pests (5.2 %),
low soil fertility, not enough rain, insufficient land, weeds and diseases (<5 %).

COMPARISON WITH SURVEY RESULTS FROM NIGERIA

In order to make inferences regarding the adoption of soybean in other parts of
sub-Saharan Africa, it is useful to compare adoption trends in other culturally
distinct areas of sub-Saharan Africa where soybean has also been introduced.
Soybean production by small-scale farmers in sub-Saharan Africa has been
largely undocumented. Only in the case of Nigeria, was sufficient data found to
permit a comparison with production practices in Zaire.


Journal For Farming Systems Research-Extension






ADOPTION OF SOYBEAN ZAIRE AND NIGERIA


Background
Results similar to those obtained in Gandajika were obtained in
surveys conducted in Nigeria by Woodworth (Woodworth et al., 1992; Smith et
al., 1993), Knipscheer et al. (1985), J. Smith (unpublished manuscript2) and
Pfeiffer (forthcoming). Woodworth conducted surveys in Benue State, located
in Southern Guinea savanna (Keay, 1959) in the eastern "Middle Belt" region
of Nigeria. Knipscheer et al. conducted their survey in Benue and neighboring
Kaduna State. Smith and Pfeiffer conducted surveys in the forest/savanna
transition zone in Oyo State in western Nigeria. Both areas were located at
approximately 70 N latitude in environments similar to that in Gandajika.
Rainfall was approximately 1300 mm, distributed bimodally (IITA, 1985;
Shannon, 1983) and soils were predominantly coarse-textured, moderately acid
Alfisols with low exchange capacities.
In Benue State, the Tiv tribe has been growing soybean since it was
introduced in 1937 (Root et al., 1987). In Ayepe and Igangan, Oyo State,
Nigeria, soybean was first introduced in 1986 and 1987, respectively (J. Smith,
unpublished manuscript2; Pfeiffer, forthcoming). Although some reference will
be made to all the locations, most of the information available on cropping
systems was obtained in Benue State, the area with the longest history of
soybean cultivation in Nigeria and where soybean is most extensively grown.

Yield
In Benue State, grain yield samples measured by Woodworth (1990)
in 40 farmers' fields in 1989 averaged 819 kg/ha and ranged from 144 kg/ha to
1609 kg/ha. The same farmers obtained average farm-level yields of 621 kg/ha
(Woodworth et al., 1992). The difference between farm level and sample yields
was attributed to shattering losses, but may also reflect field heterogeneity.
Based on interviews of 70 farmers from villages selected at random from the
eastern (Tiv) part of the state, Smith et al. (1993) estimated average soybean
yield in 1989 at 700 kg/ha.
These average yields were higher than that obtained by Gandajika
farmers. This may be partly because Benue farmers applied fertilizer to
soybean whenever possible, whereas Gandajika farmers did not. Also, many
Gandajika farmers planted on degraded soils near the homestead, thus lowering
the average yield.





2 Improvement of maize-based cropping systems. Unpublished Annual Report
for 1989. Resource and Crop Management Program. International Institute of
Tropical Agriculture, Ibadan, Nigeria.


Vol 5, No.2, 1995






D.A. SHANNON, K.M. MWAMBA, M. KUBENGU, M.C.MPOY


Area Cultivated
Woodworth (1990) collected data from 20 farmers within 40 km of
Gboko, a major center for soybean cultivation in Benue State, and 20 farmers
living more than 40 km from Gboko. Near Gboko, average area per farmer was
1.13 ha. Fifty percent of the farmers cultivated more than 1 ha, while 40 %
cultivated between 0.4 and 1 ha. Away from Gboko, average area was 0.53 ha.
Only 15 % of those sampled cultivated more than 1 ha while 45 % cultivated
less than 0.4 ha.
The average areas planted to soybean were considerably more than in
Gandajika. However, many Gandajika farmers produced two crops of soybean
each year, while Benue farmers produced only one crop.
At Ayepe, the crop was grown uniquely for local consumption
(Pfeiffer, forthcoming). Plot sizes averaged only 0.03 ha (J. Smith,
unpublished document2). At Igangan, Oyo State, the crop was sold to a
commercial buyer (Pfeiffer, forthcoming). Plot sizes were larger, some
reaching 0.5 2 ha.

Rotations and Associations
In Benue, soybean followed fallow, peanut, soybean or yam (Dioscorea
spp.) in the rotation (Woodworth, unpublished manuscript3), in southern
Kaduna State, soybean followed sorghum (Sorghum bicolor) (Knipscheer et al.,
1985), while soybean generally followed maize in Gandajika (Figure 3).
Contrary to Gandajika, only 13 % of Benue farmers grew soybean as a sole
crop. The majority intercropped with sorghum (75 %), while the remainder
intercropped with cassava (Manihot esculenta) or strip planted in citrus (Citrus
spp.) orchards. In the sorghum/soybean intercrop, sorghum grew much slower
than did soybean and reached its full height only after soybean reached
maturity. Consequently, there was little competition to the soybean crop. In
orchards, soybean was planted to suppress weeds, especially Imperata
cylindrica (A. Chichi, personal communication), and to increase land use
efficiency. Because of the open canopy of widely-spaced citrus, it may be
assumed that the soybean crop suffered only minimal shading.
In Gandajika, sorghum was rarely seen in farmers' fields, having been
almost totally replaced by maize. Some Gandajika farmers intercropped
soybean with maize, which grew faster and taller than soybean and, unless very
widely spaced, significantly shaded the soybean. Gandajika farmers therefore


3Paper presented at the National Meeting of the Nigerian Soyabean Scientists.
March 1990.


Journal For Farming Systems Research-Extension






ADOPTION OF SOYBEAN ZAIRE AND NIGERIA


did not benefit greatly from intercropping their soybean, while Benue farmers
did.

Farmer Assessment
Soil improvement was listed by 63 % of the 40 Benue farmers as a
benefit derived from growing soybean (Woodworth, unpublished manuscript3).
Advantages of intercropping were cited by 13 %. The variety grown by
Gandajika farmers, SJ 127, provided little foliage, and thus could not be
expected to contribute much to soil fertility. The advantage of this variety was
its short growing season (90-95 days) allowing two crops per year. However,
Malayan, the traditional variety grown in Benue State, was late maturing (125-
130 days), produced excessive vegetation under good growing conditions and
could be expected to suppress weeds and provide N-rich biomass.
Many of the problems listed by Benue farmers shattering losses,
labor requirements for threshing, difficulty in uprooting plants at harvest, lack
of fertilizer, and damage by insects and birds (Woodworth, unpublished
manuscript3) differed from those mentioned by Gandajika farmers. The
problems cited above may reflect the greater demand placed on labor because of
the greater areas planted and greater production in Benue than in Gandajika.
The rapid drying conditions caused by the Harmattan (dry air masses from the
Sahara desert) in Nigeria also made timely harvest more critical in Nigeria than
in Gandajika. The concern about fertilizer and insect control suggested that
Benue farmers were willing to invest to increase soybean yields.
In an earlier survey of soybean production by Knipscheer et al. (1985),
farmers in Benue State reported average seed losses of 12 % due to poor
germination, while farmers in Kaduna State reported over 50 % losses. One-
third of farmers in the two states had to replant some of their soybean crop due
to birds, seed deterioration, and in Benue State, to drought.
Poor germination was not mentioned at Gandajika. This may be
attributed in part to the fact that Gandajika farmers often planted twice each
year. As a result, the seed used for planting was stored for shorter periods, at
most 5 months in Gandajika, compared to 7-9 months in Benue State. Longer
storage under ambient conditions of high humidity and temperature are
conducive to seed deterioration and loss of viability (Dashiell et al., 1987).

CONCLUSIONS

Adoption of Soybean
The results of the surveys in Gandajika and in Nigeria disprove the
commonly-encountered belief that soybean is not suitable for production by
small-scale peasant farmers in sub-Saharan Africa. Soybean was already
integrated into cropping systems of small-scale farmers in Gandajika, and in


Vol 5, No.2, 1995






D.A. SHANNON, K.M. MWAMBA, M. KUBENGU, M.C.MPOY


Benue and Kaduna States, Nigeria. It was also being integrated into cropping
systems in Ayepe and Igangan, Oyo State, Nigeria.
Small-scale or peasant agriculture in Africa should not be viewed as
static. African farmers, like their counterparts in more developed parts of the
world, integrate new crops into existing cropping systems when it is shown to
be beneficial to them. The history of cassava, maize and peanut attest to this
fact. All were introduced into Africa around the sixteenth century (Jones,
1959; Miracle, 1966; Purseglove, 1968) and spread at different rates to become
major crops in traditional cropping systems throughout much of sub-Saharan
Africa. It is reasonable to conclude that soybean may be successfully
introduced into cropping systems in other areas of sub-Saharan Africa where
the soil and climatic conditions are suitable for growing the crop.

Integration of Soybean into Cropping Systems
The cropping practices developed by farmers in the two countries
reflect logical adaptations to local conditions. For example, the very dry
conditions during the first months of the dry season in Benue State, enable
Benue farmers to store soybean seed much longer than would be possible at
Gandajika, where relative humidity during the dry season remains high. As a
result, in Benue, a single crop of soybean is planted each year, while in
Gandajika, two crops are planted a year, despite the unfavorable harvest
conditions and labor constraints of Season A.
The popularity of intercropping in Benue and sole cropping in Zaire
reflect the difference in primary crops and their growing cycles. The fact that
Benue farmers used soybean for soil fertility improvement, while Gandajika
farmers did not, reflects the difference in growth characteristics of the soybean
varieties used.
Differences in constraints listed by farmers reflect differences in the
economic status of the crop. In Gandajika, where little commercialization of
soybean existed, farmers were concerned about sale of excess production. In
Benue, where the local marketing system was well developed, farmers were
concerned with labor constraints and lack of inputs which restricted increases
in soybean production.
Response to Markets: Plot size appeared to be influenced by whether or not
the crop was grown only for domestic consumption or for sale. Production
areas were small where soybean was grown primarily for home consumption, as
in the case of Ayepe and three quarters of Gandajika farmers. The relatively
large areas planted by farmers in Igangan and the Tiv areas of Benue State, as
well as by a minority of farmers in Gandajika, demonstrate that farmers will
expand their production of soybean when markets are available for excess
production. In later surveys conducted in Gandajika, it was learned that some
of the farmers who produced in greater quantities had obtained contracts to sell
soybean to buyers outside the local area. Thus rational management choices


Journal For Farming Systems Research-Extension






ADOPTION OF SOYBEAN ZAIRE AND NIGERIA


were made by farmers in Zaire and Nigeria, which took into account local
environmental conditions, crop varietal characteristics, the existing cropping
systems and the economic environment when determining the place of soybean
within farmers' overall cropping system.

Crop Surveys
The surveys reported here played an important role in justifying
research agendas. The survey at Gandajika showed that soybean was widely
grown and was integrated into cropping systems of the community, thus
confirming its relevance to farmers and providing justification for continued
research into the crop at a time when such research was being questioned. The
surveys in Nigeria undoubtedly played a similar role in justifying continued
research in soybean there also. In addition, the insights into farmer practices,
constraints and perspectives gained from these simple multi-disciplinary, crop-
specific surveys served as a basis for planning future research that would be
relevant to local conditions and address real constraints to production. Much
useful information was gained in a very short time and at limited cost.

Implications for Future Research
Research into new crops or cropping practices always involves a risk
that the new technology will not be appropriate or accepted by farmers.
Although results from a single location would be inconclusive, evidence of
adoption and integration into existing farming systems in areas with different
bio-physical and socio-economic conditions demonstrates that the new
technology can be widely accepted and extended into new areas. In this paper,
we have shown that soybean has been successfully integrated into farming
systems in areas of Zaire and Nigeria differing in bio-physical and socio-
economic conditions and existing cropping systems. Continued research to
improve soybean production may be justified based on the expectation that the
crop will be integrated into farming systems in other parts of sub-Saharan
Africa.

ACKNOWLEDGEMENTS

This research was financed by the United States Agency for International Development,
Project 660-0091, Applied Agronomic Research and Extension (RAV), with technical
assistance from the International Institute of Tropical Agriculture, Ibadan, Nigeria. The
authors are grateful to Director Kilumba Ndayi of the Programme National
L6gumineuses for providing staff for the survey and to Dr. Joyotee Smith and Dr. L.
Upton Hatch for useful suggestions.


Vol 5, No.2, 1995







D.A. SHANNON, K.M. MWAMBA, M. KUBENGU, M.C.MPOY


REFERENCES

Dashiell, K.E., L.L. Bello and W.R. Roo. 1987. Breeding soybeans for the tropics. Pages 3-16 in S.R.
Singh, K.O. Radie and K.E. Dashiell, eds., Soybeans for the Tropics: Research, Production and
Utilization. New York: Wiley & Sons.
IITA, 1985. Annual Reportfor 1984. International Institute of Tropical Agriculture, Ibadan, Nigeria.
IITA, 1987. IITA Annual Report and Research Highlights 1986. International Institute of Tropical
Agriculture, Ibadan, Nigeria..
IITA, 1988. IITA Strategic Plan 1989-2000. International Institute of Tropical Agriculture, Ibadan,
Nigeria.
IITA 1992. Sustainable Food Production in Sub-Saharan Africa: 1. IITA's Contributims.
International Institute of Tropical Agriculture, Ibadan, Nigeria.
Jones, W.O. 1959. Manioc in Africa. Stanford: Stanford University Press.
Keay, R.J.W. 1959. An outline of Nigerian vegetation. 2nd edition. Lagos, Nigeria: Federal
Government Printer.
Knipscheer, H.C., K.M. Menz and P. Ay. 1885. The production and market potential of soybeans in
Nigeria. Quarterly Journal of International Agriculture 24 (2): 171-184.
Miracle, M.P. 1966. Maize in Tropical Africa. Madison, WI: University of Wisconsin Press.
Pfeiffer, J.M Forthcoming. Factors affecting adoption of soybeans into cropping systems and diets by
small farmers, rural households and petty traders. Paper presented at the Symposium on
Sustainable Agriculture in Africa: Socio-Cultural, Political, and Economic Considerations. Ohio
State University Center for African Studies. 25-26 May 1990.
Purseglove, J.W. 1968. Tropical Crops:Dicotyledons. Lndon: Longmans.
Root, W.R., P.O. Oyekan and K.E. Dashiell. 1987. West and Central Africa: Nigeria sets example for
expansion of soybeans. Pages 81-85 in S.R. Singh, K.O. Rachie and K.E. Dashiell, eds., Soybeans
for the tropics: research, production and utilization. New York, NY: John Wiley & Sons
Shannon, D.A 1983. A study of factors responsible for variable growth of soybeans in the Southern
Guinea savanna of Nigeria. Unpublished Ph.D. Thesis. Comell University, Ithaca, New York.
Shannon, D.A and Mwamba, K.M. 1994. Adoption of soybean in sub-Saharan Africa: a comparative
analysis of production and utilization in Zaire and Nigeria. Agricultural Systems 46: 369-384.
Shannon, D.A, M. Ngoyi and N. Kilumba. 1988. On-station agronomic trials with soybeans and
groundnuts. Pages 163-165 in International Institute of Tropical Agriculture. Resource and Crop
Management Program Annual Report for 1986. Ibadan, Nigeria.
Smith, J., J.B. Woodworth and K.E. Dashiell. 1993. Government policy and farm level technologies:
the expansion of soybean production in Benue State, Nigeria. Agricultural Systems in Africa 3
(1): in press.
Vanhamme, J., J. Clement, and A. Collin. 1955. Aperqu sur l'Fconomie agricole de la Province de
Kasai. Direction de 1'Agriculture des Forets et de lIlevage, Ministre des Colonies, Royaume de
Belgique, Bruxelles.
Vanneste, G. 1986. Document IV La culture du soya. Pages 171-172b in Seminar National sur le
Soja. Proceedings of a conference held at Kananga, Zaire, 3-10 May 1972. Comit6 de
Coordination pour le Developpement en Ripublique du Zaire and Comit6 Diocesain de
Developpement Integral de Kananga. R-edition.
Weingartner, K.E., K.E. Dashiell and AI. Nelson. 1787. Soybean utilization in Africa: making a place
for a new food. Food and Nutrition 13: 21-28.
Whingwiri, E.E. 1987. Soyabean produdion in communal and small-scale sectors. pp. 13-17 in
Soyabeans in Southern Africa. Proceedings of a workshop for the Southem Africa region on the
basics of soyabean cultivation and utilization. University of Zimbabwe, Harare, Zimbabwe.
Woodworth, J., J. Smith and K. Dashiell. 1992. Survey and crop season study of soybean in Benue
State Nigeria. Tropical Oilseeds Journal. 1: 75-76.


Journal For Farming Systems Research-Extension










THE IMPACT OF PLAN SIERRA'S OUTREACH
PROGRAMS ON THE SUSTAINABILITY OF
HILLSIDE FARMING IN THE DOMINICAN
REPUBLIC

Scott G. Witter' andMichael P. Robotham2

ABSTRACT

This paper focuses on the iinpact that Plan Sierra's
agricultural outreach programs have had on the sustainability
of small Dominican hillside farms in a 1,700-square-
kilometer region. Comparisons are made between farms
adopting controlled innovations and those who have not had
any contact with Plan Sierra and represent the state of
farming systems in the region if Plan Sierra did not exist.


INTRODUCTION


The analysis presented in this paper is based on Harrington's (1992) article on
alternative approaches to measuring sustainability as an objective of research.
Harrington describes three major approaches to sustainability: agroecology,
equity, and sustainable growth. He also provides a classification of sustain-
ability problems based on internal vs. external, reversible vs. irreversible, and
agricultural productivity vs. public health variables. And Harrington concludes
that such measurements are in their early stages and much remains to be done
by FSRE practitioners in partnership with disciplinary specialists.
The objective of this research is to measure the impact of Plan Sierra's
outreach innovations on the sustainability of small hillside farming systems in
the Dominican Republic. To measure their impact, comparisons are made
between two kinds of farming systems within similar agroecological zones-(1)
farms that have been modified based on external contact with Plan Sierra and
(2) those that have evolved based on internal inputs from family and friends.


1
Associate Professor, Department of Resource Development, Michigan State University.
2
Dept. of Agriculture and Soil Scimnce, University of Hawaii-Manoa.


VoL 5, No. 2, 1995






WrITER AND ROBOTHAM


Farming systems that have been modified as a result of a series of
innovations delivered by Plan Sierra are considered controlled systems. Those
farms that have not had contact with Plan Sierra are called state because they
represent the state of Dominican hillside farmers if Plan Sierra did not exist.
Harrington (1992) uses the example of the effect of tillage practices (a control
innovation) on the state of soil depth. A cause and effect relationship exists
between the use of control innovations and the condition of the existing
farming system.
From an agroecology perspective the state system represents the status of
the farming system without directed innovations (outreach and training
programs by Plan Sierra), and any changes or alterations in practice were
introduced by an internal mechanism, such as friends or family. In the
controlled system, the inputs and practices were initially external to the
community and were introduced by Plan Sierra and not by an internal
mechanism.
When a farming system reaches its production goals and still maintains the
quality of its natural resource base, it is said to be sustainable. Such a system
extracts only as many resources as the environmental system can renew,
produces fewer wastes than can be absorbed by the environment, and optimizes
the efficiency of renewable resource use (Conway, 1994; Harrington, 1992;
Barbier and McCracken, 1988).
This paper focuses on the internal questions related to decision making by
the farmers to work with Plan Sierra, adopt conservation practices, use
chemical inputs, maintain a given crop and animal diversity, produce crops and
animals for sale, hire off-farm labor. The paper also details the migration and
family characteristics of the farmers questioned. Directional measurements are
made using Cramer's V and frequency statistics to illustrate the degree of
difference between the control and state systems (Craft, 1991). To be
statistically significant the CV values have to be significant at the p=.05 level.
Cramer's V Strength of associations are measured as follows: < .10 weak, .11-
.25 weak to moderate, .26-.40 moderate, .41-.50 moderate to strong, and > .50
strong (Craft, 1991).


PLAN SIERRA

Plan Sierra is located on the north central slope of the Cordillera Central
overlooking the agriculturally rich Cibao Valley. The region has a range of
micro climates allowing for a wide variety of crops to be grown (i.e., onions to
coffee). The project area covers 1,700 square kilometers.


Journalfor Farming Systems Research-Extension






IMPACT OF PLAN SIERRA ON SUSTAINABILrTY OF HILLSIDE FARMING


The most current population estimate comes from the 1980 census, which
indicated that 110,000 people lived in the region. Ninety percent, virtually all
farming families, lived in the countryside in a series of small villages; 10
percent nonfarmerss) were reported as living in three small cities. The 90
percent living in the country represent Plan Sierra's primary clientele.
Plan Sierra was created by a grant from the government of the Dominican
Republic in 1979 to address the needs of the rural poor. Its objectives were and
are to (1) create a demonstration project for managing upland and mountainous
agriculture; (2) develop a coordinating mechanism to link existing
management institutions together to address the problems of the sierra; and (3)
respond in a timely manner to small farmer needs rather than to pursue
predetermined targets. To this end, the core of all programs at Plan Sierra are
the small farmers and their rural communities (Santos and Quezada, c1979).
Plan Sierra was officially removed from all government control in 1983,
although it was still receiving almost 70 percent of its operating capital from
the government (de Janvry and Hecht, 1984). By 1991 this figure had
decreased to 18 percent and is currently under 10 percent of the total yearly
operating budget. The rest comes from U.S. foundations, the German
Development Assistance Program (GTZ), income-generation projects, grants
from Dominican foundations, and private donations.
Over the course of its operations, Plan Sierra has conducted activities in the
areas of infrastructure development, health, education, and agriculture (de
Janvry and Hecht, 1984). The reduction of soil erosion through better farm
management and the proper use of conservation practices has been a central
focus of the agricultural program (Jimenez and Gutierrez, 1993). Extension
practitioners-the outreach workers of Plan Sierra-have used information
pamphlets, radio programs, community meetings, site visits, and intensive
training at the Kellogg Foundation-sponsored Los Montones training facility.


FARMING SYSTEMS ANALYSIS

For this study, a farming system was defined as "a unique and reasonably stable
arrangement of farming enterprises that the farm household manages according
to well-defined practices in response to the physical, biological, and
socioeconomic environments and in accordance with the household's goals,
preferences, and resources" (Shaner et al., 1982). This definition stresses that a
farming system exists within an agroecological zone and responds to the
physical, biological, and socioeconomic environments. The importance of this
context is also stressed by a number of other authors (Brush and Turner, 1987;
FAO, 1989; Friedreich, 1992).


Vol 5, No2, 1995






WrITER AND ROBOTHAM


DATA COLLECTION

A questionnaire was developed in concert with the Instituto Superior de
Agriculture (ISA), the Center for Urban and Rural Development (CEUR) at
Universidad Catholic Madre y Maestra (UCMM), and Plan Sierra staff in
Santiago, Dominican Republic. The questionnaire was designed to collect a
wide array of information regarding the farmstead: agroforestry, soil and water
conservation management practices, farm inputs and outputs, credit, use and
availability of water, demographics, inheritance, and coffee production. The
survey was field tested twice during the fall and winter of 1992-93 by faculty
from ISA. During March 1993, a group of three trained ISA enumerators, who
had not had previous contact with Plan Sierra, and a research assistant
(Michael Robotham) from Michigan State University conducted the survey.
Interviews where held in Spanish by the Dominican enumerators.
Participants in the study were selected by the interviewers using a stratified
random sampling procedure. Enumerators began at the center of a village and
proceeded out in three different transects following existing roads and trails
leading to the farms. Each was provided with a list of farmers living in the
community who had participated in Plan Sierra's outreach programs. Upon
locating a controlled farm, the enumerator asked to speak to the head of the
household (in all cases a man) and interviewed that person. The enumerator
then would proceed to the nearest neighbor who had not taken part in any of
Plan Sierra's outreach programs and interview that individual. This approach
was taken to guarantee a direct comparison of control and state farms within
same community and agroecological zone.


MEASURING THE DIFFERENCES IN CONTROLLED
AND STATE SYSTEMS

Seven groups of questions were selected to measure if there were statistically
significant differences between the controlled and state farms. The questions
dealt with: (1) farmer and farmstead characteristics; (2) adoption and use of
soil and water conservation practices; (3) migration and family characteristics;
(4) hiring of off-farm labor; (5) crop and livestock diversity; (6) commercial
sale of crops and animals; and (7) use of chemical inputs.

Farmer and Farmstead Characteristics
Comparisons were made between farmers and farmsteads of the two groups
to discover whether there were significant differences between the type of
farmers (or farm) that Plan Sierra worked with and those they did not. To
determine significant differences (p 2 .05), the controlled (n=96) farms data
were separated from the state (n=65). The size of farms ranged from 1 tarea (16


Journalfor Farming Systems Research-Extension






IMPACT OF PLAN SIERRA ON SUSTAINABILrrTY OF HILLSIDE FARMING


tareas=l hectare) to 500. Ninety-two percent of the controlled group's farms
averaged 1.92 hectares and 98.5 percent of the state farms averaged 2.06
hectares. Cramer's V (CV) tests of all farms indicated that there were no
statistically significant (CV=.571, p=.573) differences between the two groups
using size of the farm as a variable.
Further comparisons of the number of years farming, number of land
holdings, number of farmers who owned their farms, and the age of the
farmers yielded no statistically significant difference between the groups. A
significant difference was recorded for income; however, 61 percent of the
control and 78 percent of the state groups chose not to respond to questions
regarding their income.
Farmers where asked to rank the level of importance of obtaining farming
inputs such as seeds, fertilizer, herbicides, pesticides, and credit. Again, no
statistically significant differences were found. Farmers were also asked to
rank the importance of environmental problems to their farm, specifically soil
erosion, flooding, drought, insects, and weeds. A moderately significant
difference was established between the two groups with respect to soil erosion (
CV=.266, p=.010). Eighty percent of the control group ranked soil erosion as
very important or important compared to 64 percent of the state group. No
other significant differences were recorded.

Adoption and Use of Soil and Water Conservation Practices
The adoption of conservation practices was seen as a good indicator of the
farmer's willingness to adopt an important innovation. Use and long-term
maintenance of conservation practices were seen as measures of the farmer's
understanding of the importance of long-term stability of the farming unit.
The strongest and most significant differences between the control and state
groups were recorded for questions related to the adoption and maintenance of
soil and water conservation practices (SWCP). Questions about SWCPs were
divided into three groups related to the length of adoption and types of practices
used, where the farmer first learned about SWCPs, and who provided the most
useful information on SWCPs.
Ninety-five percent of the control group and 25 percent of the state. group
had adopted SWCPs. Cramer's V tests indicated a very strong and highly
significant difference (CV=.761, p=.000). Thirty-seven percent of the control
and none of the state participants used SWCPs on all of their farm land
(CV=.836, p=.000). Comparisons of the length of time SWCPs have been
practice were also strongly and significantly different (CV=.781, p=.000) with
the control participants adopting and maintaining longer than state. Only one
farmer who had adopted SWCPs reported ceasing to use them, and then only
because he was 80 years old.


Vol 5, No2, 1995






WITTER AND ROBOTHAM


Only 5 percent of the control farmers indicated family and friends as
sources of SWCP information. By contrast, all of the state group who practiced
SWCP listed family and friends as their source of SWCP information. When
asked which source of information was most useful, 88 percent of the control
group indicated the external input from Plan Sierra. When the state farmers
were asked for their most useful sources of SWCP information, only 15 percent
of the farmers interviewed felt that they had received useful information from
family and friends.

Migration and Family Size Characteristics
The size of the family and the number of individuals who had to migrate
from the farm were seen as good indicators of the stability of the farming unit
to meet the needs of the family unit. Analysis regarding the demographics and
migration of the two groups showed little significant difference, except for the
number of adults living at each farm. The control farms averaged 3.8 adults,
while the state averaged 3.1 (weak-moderate difference, CV=.208, p=.003).
The control group average 2.2 children per farm and the state 2.4, and there
were no significant differences between the two groups with respect to the ages
of the farmers and their children.
There were weak to moderate differences in the number of children per
farm who had migrated to the United States and to other parts of the
Dominican Republic (CV=.166, p=.107; and CV=.190, p=.055); slightly more
(2 and 10 percent) came from the controlled farms. And no significant
difference existed in the number of adults who migrated from the farm.

Hiring Off-farm Labor
The ability to hire and pay off-farm labor was deemed to be a good indicator
of a prosperous farming unit. When asked if their farm required outside labor,
73 percent of the control and 66 percent of the state farmers answered yes.
Significant differences were recorded in the amount of labor needed for
weeding (CV=.275, p=.002; control 56 percent, state 29 percent) and
harvesting (CV=.194, p=.047; control 53 percent, state 58 percent). No
significant differences were recorded for soil preparation or planting, but in
both instances a slightly higher percentage of labor was needed on control
farms.

Crop and Livestock Diversity
Crop diversity and livestock diversity were measured as the number of
different types and species found on each farm, respectively. The assumption
was made that the more diverse the farm, the more stable the farm would be to
environmental and market fluctuations. A moderate to strong difference at a
significance level of. 148 was recorded for crops.


Journalfor Farming Systems Research-Extension






IMPACT OF PLAN SIERRA ON SUSTAINABLITY OF HILLSIDE FARMING


Commercial Sale of Crops and Animals
The demonstrated ability to produce enough crops and/or animals to permit
a portion to be sold was assumed to be a good measure of the stability and
prosperity of the farming system type. Once again, a moderate to strong
difference was recorded at the .126 significance level below the .05 level
deemed acceptable for this study.

Use of Chemical Inputs
The ability to purchase and use insecticides, herbicides, and commercial
fertilizers was assumed to be a relevant measure both of the farmer's
willingness to accept new innovations and of a farmstead that could produce
enough to allow the farmer to purchase these inputs. Comparisons indicated a
highly significant moderate difference (CV=.398, p= .000) between the control
and state. Sixty-eight percent of the control farmers and 28 percent of the state
group used commercial fertilizer.
A highly significant weak-to-moderate difference (CV=.191, p=.015) was
recorded for use of insecticides; 41 percent of the control farmers and 22
percent of the state farmers used insecticides. There was no significant
difference between the two groups for the use of herbicides with only 3 percent
of the control and 5 percent of the state farmers using this input.


CONCLUSIONS

The objective of this research was to measure the impact of Plan Sierra's
outreach innovations on the sustainability of the small hillside farming systems
found in a 1,700-square-kilometer area of the Dominican Republic. To do this
farmers were divided into two groups: the control group consisted of farmers
who were affected by the external influences of outreach programs sponsored by
Plan Sierra; the state group consisted of farmers who relied on internal
influences from family and friends. Comparisons of seven variables were made
between the groups of farmers living in the same agroecological zone: (1)
farmer and farmstead characteristics, (2) adoption and use of soil and water
conservation practices, (3) migration and family characteristics, (4) need for
off-farm labor, (5) crop and livestock diversity, (6) commercial sale of crops
and animals, and (7) use of chemical inputs.

Farmer and Farmstead Characteristics
Comparisons were made between the control and state groups to see if there
were significant differences in farm size, ownership, use of credit, number
years farming, and income. No significant differences were determined


Vol 5, No2, 1995






WrITER AND ROBOTHAM


between the groups for farm size, ownership, and number of years farming.
There were also no significant differences in farmers' perception of the
difficulties in obtaining a variety of farming inputs, environmental problems,
and pests. These combined results indicated that there was no specific type of
farmer or farm that Plan Sierra was targeting for its programs and, thus, the
two groups could be logically compared.
The only significant differences recorded in this category were income, the
use of credit, and the perception that drought, soil erosion, selling the harvest,
and weeds were problems; more control than state farmers answered yes in
each instance. Further investigation of the control group showed that many of
the farmers reporting higher incomes and using credit were involved in coffee
production. Plan Sierra encourages coffee production for erosion control and
increased income. Because of Plan Sierra's programs control farmers also put
a higher emphasis on soil erosion and weeding.
A higher percentage of state farmers found it more difficult to obtain seeds
and sell their harvest, and felt that drought was a problem. The lack of an
organized training program-other than Plan Sierra-helping farmers to raise
seeds and find a market for their products indicates that for state farmers there
are no programs that are adequate substitutes for those implemented by Plan
Sierra.

Adoption and Use of Soil and Water Conservation Practices
Plan Sierra has had a dramatic impact on the adoption and maintenance of
soil and water conservation practices. This has been accomplished through the
use of demonstration farms and by onsite extension practitioners who stress
long-term thinking and use a number of self-help innovations rather than cash
payments to gain farmer adoption. Plan Sierra is having a positive impact on
the local environment, as well as on downstream ecosystems, by reducing
sediment loads in the rivers and streams that flow through the region.

Migration and Family Size Characteristics
Significantly more adults were found to be living in the control families
than state ones. In an attempt to explain this, comparisons were made between
the groups to see if more control farmers were inheriting (staying on the family
farm) or buying new farm land. Fifty-five percent of the control farmers
indicated that they inherited their primary farm land compared to 43 percent of
the state group. This would appear to be an indicator that more control farmers
are taking over the family farm, which, in turn, is an indicator that the farming
system itself is more environmentally sustainable. Additional field research is
required to confirm this assumption.


Journal for Farming Systems Research-Extension






IMPACT OF PLAN SIERRA ON SUSTAINABILITY OF HILLSIDE FARMING


Only weak to moderate differences were recorded for migration, where
people migrated, and the number of family living at home-slightly more
children from the control farms were recorded as leaving the region. Open-
ended questions relating to migration found the most often-given reasons for
leaving were related to school, employment, and moving to live with relatives.
This information indicated that there was no clear difference between the two
groups and that Plan Sierra's external outreach programs were not clearly
influencing the migration of children from the farm.

Hiring Off-farm Labor
Control farms were more likely to need off-farm labor for soil preparation,
planting, and weeding than state farms. Plan Sierra emphasis on SWCPs
means that more time is required in the field for contour plowing and plowing
between live or dead barriers and on terraces. Weeding is considered an
essential part of water conservation and is stressed during training sessions.
The analysis indicates that there is little difference in the amount of labor
needed for harvesting.

Crop and Livestock Diversity
While not significant at the .05 level, there was a moderate-to-strong
difference between the groups in the number of crops grown. Plan Sierra trains
farmers to raise their own seeds, often providing the first season's seeds to those
who participate in its programs. The farmers are then required to pay Plan
Sierra back in kind during the next year. This provides stability to the system
and allows Plan Sierra to maintain a seed stock for future participants. A
similar system is used for animals. The strong significant difference for animal
diversity appears to be more reflective of agroecological factors than Plan
Sierra's programs.

Crops and Animals for Sale
Though not significant at the .05 level, the same pattern exists for the
farmers who have crops and animals for sale as for those who have a higher
degree of diversity in their farms. Plan Sierra provides information and help to
the farmers for marketing their crops, especially for coffee growers. This
appears to be beginning to have an impact, but will require further study.

Commercial Chemicals Used
The strongest, most significant difference between the groups was in the use
of commercial fertilizers. Plan Sierra has a program that provides its
participants with more ready access to fertilizer and encourages its use.
Currently, there are no specific data available to support the contention that
this increased fertilizer use is having a significant impact on the yield of the
crops produced. Nevertheless, it is a reasonably safe assumption that increased


Vol 5, No2, 1995







WITTER AND ROBOTHAM


use of fertilizers does increase crop yield when all other inputs are held
constant.
The research presented in this paper indicates that Plan Sierra has had a
positive long-term impact on the region and farmers living within boundaries
of the project. This is especially true of the adoption and maintenance of soil
and water conservation practices. The long-term maintenance of these
conservation practices exceeds any of those previously documented programs in
the Dominican Republic (Carrasco and Witter, 1993). Plan Sierra has
demonstrated that long-term sustainability of physical resources can be
accomplished through well developed and presented outreach programs.
Because of the success of these conservation programs and their community
organizational programs, Plan Sierra's outreach programs are being
incorporated into the current Rio Yaque del Norte Watershed Project (Witter
and Carrasco, 1996). Plan Sierra is located in the headwaters of the Rio Yaque
watershed. Plan Sierra's outreach providers and participating community
leaders have been sought out to become involved in designing new programs
for the Rio Yaque project. There is still much that we can learn about the
organization of women's groups, participant-paid youth environmental
programs, and community health programs used by Plan Sierra.



REFERENCES


Barbier, E., and J. McCracken. 1988. Glossary of selected terms in sustainable economic
development. IIED Gatekeeper Series SA&. International Institute for Environment and
Development.
Brush, S.B., and B.L. Turner Jr. 1987. The nature of farming systems and views of their change. In.
B.L. Turner Jr. and S.B. Brush, eds., Comparative farming systems. New York: The Guilford
Press.
Carrasco, D.A. and S.G. Witter. 1993. Constraints to sustainable soil and water conservation: A
Dominican Republic Example. Journal ofthe Human Environment 22(6): 347-350.
Conway, G.R. 1994. Sustainability in agricultural development: Trade-offs between productivity,
stability, and equitability. Journal of Farming Systems Research-Extension 4(2): 1-14.
Craft, J.L. 1991. Statistics and analysis for social workers. Itasca, IL: F.E. Peacock Publishers.
de Janvry, A., and S. Hecht. 1984. Report de la evaluacion de los primeros cinco anos del Plan
Sierra. Unpublished evaluation, Plan Sierra, San Jose de Las Matas. Dominican Republic.
Food and Agriculture Organization (FAO). 1989. Farming systems development: Concepts, methods,
and applications. Rome: FAO.
Friedreich, K.H. 1992. A farming systems approach to agricultural development. In Readings in
farming systems. Rome: FAO, pp. 107-114.
Harrington, R.R. 1992. Measuring sustainability: Issues and alternatives. Journal of Farming
Systems Research-Extension 3(1): 1:22.
Jimenez, A., and L. Gutierrez 1993. Notes from a presentation and discussion with Jimenez and
Gutierrez at Plan Sierra on 10 March 1993.
Santos, B., and N. Quezada. c1979. Report on Plan Sierra. Institute Superior de Agricultura, La Herra,
Dominican Republic.


Journalfor Farming Systems Research-Extension







IMPACT OF PLAN SIERRA ON SUSTAINABILITY OF HILLSIDE FARMING


Shaner, W.W., P.F. Philipp, and W.R. Schmehl. 1982. Farming systems research and development:
Guidelinesfor developing countries. Boulder, CO: Westview Press.
Witter, S.G. and D. A. Carrasco. 1996. Water quality: a development bomb waiting to explode: a
Dominican example and possible solution. Journal of the Human Environment 25(3): 199-204.


Vol 5, No2, 1995










AN INTERROGATIVE APPROACH TO SYSTEM
DIAGNOSIS: AN INVITATION TO THE DANCE



P.G. Cox, A.D. Shulman, P.E. Ridge, M.A. Foale, andA.L. Garside'


ABSTRACT

We collected the rules that farmers assert they use in their
decisions to implement various technological components in
broadacre dryland farming systems in Central Queensland.
Coherent rule sets were obtained for the use of: fertilizer;
opportunity cropping; surface management; and crop
sequence. But these rule sets did not have a linear tree-like
structure. In some situations, rules are not organised in
sequences of actions but are triggered by external events.
There were interconnections between the different sets of
rules. The rule sets captured differences between farming
systems, and between regions with a different history of
agricultural development. They were stable in response to
perturbation (e.g. severe drought), and adaptive to long-term
trends in system status (e.g. fertility decline). The rule sets are
an effective and efficient way for farmers to manage the use
of technological components in a variable environment.

Both exploratory Rapid Rural Appraisal (RRA) and
ethnographic decision tree modelling are inadequate
qualitative approaches for system diagnosis in face of the
sequential, inter-dependent and contingent nature of decisions
about the use of technological and other options. The rule sets
do provide useful models of farmer decision-making. They
help to specify: criteria for the design of revised models; a
baseline by which to judge the claims of new models; a source
of insights; entry points for change; and the basis for
substantial dialogue. We have adopted a dance metaphor, in
which researchers are invited to learn the rules of complex,

' The first, third, and fourth listed authors are with the Agricultural Production Systems Research Unit,
PO Box 102, Toowoomba, Queensland 4350, Australia; the second author is with the Communication
Research Institute of Australia, PO Box 8, Hackett, ACT 2602, Australia; and the fifth author is with
the Queensland Department of Primary Industries, PO Box 6014, Rockhampton, Queensland, 4702,
Australia.


Vo 5, No. 2, 1995






Cox, SHULMAN, RIDGE, FOALE, AND GARSIDE


dynamic and inter-linked dances, to help both explore and act
on these issues. The dance is a more appropriate metaphor
than a tree for articulating change in the way in which
researchers and farmers communicate.

INTRODUCTION

Rapid appraisal methods (Doorman, 1991; Kumar, 1993; van Nieuwkoop et al.,
1994) have been a feature of farming systems research (FSR), since the early
1980s. Collinson (1981) saw this as a low cost approach for understanding the
issues faced by small farmers. Hildebrand (1981) emphasised the value of
bringing multiple perspectives to bear on the identification of problems,
constraints and opportunities defining the performance of agricultural
production systems i.e. for system diagnosis. By the mid-1980s, a rapid rural
appraisal (RRA) was seen by many as "the essential first step" in FSR (Beebe,
1985). The late 1980s saw the development of a wide range of rapid appraisal
methods for various purposes, including the instigation and management of'
change in agricultural systems (participatory rural appraisal or PRA) e.g.
McCracken et al. (1988), Feldstein and Jiggins (1994). RRAs are increasingly
used as an active intervention to promote change in directions owned by certain
(usually inadequately empowered) constituents. They are best for indicating
problems or opportunities to be explored further, and for initiating a process of
exploration and change, rather than for making generalisations.
Rapid appraisal methods have been used in Australia by researchers
including Ison and Ampt (1992), Dunn and Gray (1992), and Hamilton (1991).
In Central Queensland, we also used the concept of a participatory rural
appraisal to help identify issues requiring attention in crop production, and to
initiate a research programme to address them (Cox et al., 1993a). There was
sufficient agreement between the farmers involved, and between the farmers
and the researchers, about the issues (largely to do with the management of
nitrogen in a region only recently beginning to experience noticeable decline in
native nitrogen fertility) to form the basis for a research programme.
Gladwin (1976, 1983, 1989) presented her ethnographic decision tree
methodology in part to counter some perceived inadequacies of these rapid
methods for system diagnosis. In general, descriptive/exploratory RRA has
been successful at identifying broad themes requiring the attention of
development workers; a PRA provides a framework for beginning to address
those issues. However, we are given little guidance about how to proceed in
order to pinpoint opportunities for intervention once issues have been identified
although, ultimately, these are matters for negotiation between stakeholders.
There is insufficient depth to the understanding provided by an exploratory
RRA adequately to specify a starting point for negotiations; nor for the detailed
design of an intervention to address specific issues; nor for assessing the


Journal for Farming Systems Research-Extension






AN INTERROGATIVE APPROACH TO SYSTEM DIAGNOSIS INVITATION TO THE DANCE 69

adequacy of any intervention in satisfying the design criteria established during
the process. Ethnographic decision tree modelling potentially provides a
formalism to do this. Hierarchical (ethnographic) decision models of farmer
behaviour have been used in Australia by Murray-Prior (1994).
Our own interest was in the possibility of specifying opportunities for
decision support for broadacre dryland crop production in southern and central
Queensland, Australia, through better understanding of decision-making
processes and the decision-making environment of farmers (Cox et al., 1993b).
The Agricultural Production Systems Research Unit is a simulation modelling
group established in 1991 jointly by the Queensland Department of Primary
Industries and the CSIRO Division of Tropical Crops and Pastures in the belief
that decision support technologies, based on the simulation of biophysical
processes, can contribute to improved farm-level decision-making.
We began by attempting to represent the content of certain production
decisions by specifying the relationships between physical components (such as
weather, soil water and nitrogen supply) and actions. We proposed to use
Gladwin's decision tree methodology to help uncover the basis for decisions
about the use of individual technological components such as nitrogen fertilizer.
Decisions are patterned into event nodes and decision nodes which are explored
in a logical sequence. This is a behavioral decision tree, comparable in
representation (but not in normative content) with the representations of
decision analysis (e.g. Goodwin and Wright, 1991). The models can be
validated against external data sets, and used predictively.
The attractiveness of Gladwin's framework lay partly in its apparent ability
to expose an underlying structure (of linkages between events, and weaknesses
in the response of farmers to them) that we might usefully exploit to specify
better the design of interventions to support decision making about the use of
various technological components. Also, the superficial isomorphism between
an ethnographic decision tree model and a rule-based expert system (Alty and
Coombs, 1984) suggested that this might be a representation that is adequate
not only to identify entry points for change but also to effect change. However,
we found this approach inadequate for our purpose. It became clear that an
alternative metaphor was needed. This paper explains why and presents an
alternative and, we believe, more productive metaphor for system diagnosis.

METHODS

During 1992/93, we interviewed ten farmers in Central Queensland (four in
Banana near Moura in the Dawson-Callide region, six in Capella in the Central
Highlands) about the way in which they use, or do not use, several
technological components believed by researchers to contribute to a more
sustainable farming system: opportunity cropping (growing a crop whenever a
planting opportunity occurs following rainfall); use of inorganic fertilizer (as


Vol 5, No. 2, 1995






Cox, SHULMAN, RIDGE, FOALE, AND GARSIDE


the natural fertility of the land declines to a point where fertilizer application
becomes profitable); purposeful sequences of crops (e.g. in rotations including
legumes and a pasture phase); and reduced tillage (to improve water
infiltration, reduce run-off and erosion, and increase cropping intensity). The
preoccupation of researchers with these particular technological components
reflected their appreciation of the importance of water and nitrogen as
constraints on the performance of local farming systems. The study period was
characterized by a prolonged drought.
The farmers had already begun working with us on a pilot programme of
on-farm research (Cox et al., 1993a) and relationships between researchers and
farmers had been established before the interviews. The Banana group was a
natural group that was operating effectively before our intervention in the on-
farm programme; the Capella group was artificial. They were recognized
locally as amongst the best farmers in the two areas. Their approach to farming
captures many aspects of best current practices. The original interviews were
conducted by an agronomist and an economist in the way described by
Collinson (1981). Conversations were directed towards discussion of the
technological components of interest to researchers, but ranged widely within
that constraint. A questionnaire was not used; nor were the interviews tape
recorded. Emphasis was on how farmers describe their current practices, and
interrogatively exploring opportunities for improved agricultural production.
The interviews lasted between two and six hours.
Following the initial round of interviews, we patterned what we had been
told as a series of production rules (i.e. IF...THEN). We made no attempt to
distinguish what people said they do and what they do do (Argyris et al., 1985).
The rules were grouped into sets corresponding with the technological
components we were interested in. The responses were checked and adjusted in
a second round of interviews with individual farmers, and in a third round of
group discussions at the two main locations. We used copies of the edited notes
to feed back to the farmers on both occasions. We also interviewed four other
farmers in the Banana area whom our original respondents suggested we
should contact to get further insight into local decision-making behavior. Some
information about other technological components was also documented. Later,
we tried to represent these rules as decision trees as suggested by Gladwin. A
preliminary report of this activity has been presented previously (Cox et al.,
1993b).
RESULTS

We were only partly successful in terms of establishing a coherent tree-like set
of logically related rules that govern behavior. For instance, we were able to
impose some structure in the form of a partial decision tree describing the rules
for winter crop planting and fertilizer use by farmers in Central Queensland,
Figure 1. In this case we could identify more depth, in terms of the number of


Journalfor Farming Systems Research-Extension






AN INTERROGATIVE APPROACH TO SYSTEM DIAGNOSIS INVITATION TO THE DANCE 71

explicitly-stated event and decision nodes and the relationships between them,
than for most of the decisions we looked at, although this depth was not
apparent in our initial patterning of the rule sets elicited from farmers.

Differences Between Farmers/Farming System
Differences between farmers/farming systems were most pronounced in the
Banana area where the rules lor all four components were clearly different for a
producer who was primarily interested in cattle (and grew crops for feed),
through mixed cattle/cropping enterprises, to a cropping enterprise where cattle
are agisted following crop harvest until it rains. For example, the farmers with
a cattle with crops system used a different pattern of crop sequence (typically
sorghum-sorghum-sorghum-sorghum-oats) than those with crops with cattle
and crops systems. The latter plant anything suitable depending on the time of
year when an opportunity arose, provided that they have sufficient reserves of
soil water and a planting opportunity occurs ("opportunity cropping"). The
cattle with crops system manager was not concerned about soil compaction
from cattle; the crops system manager was clearly concerned about this and
took measures to avoid it happening. The resolution of the rule sets was thus
sufficient to distinguish different farming systems, Table 1.
If the farmer is mainly interested in cattle, the crop sequence is dominated
by grain sorghum. This can be grown more than three years in a row including
ratoon crops. No paddocks are exclusively reserved for crops. If the main
interest is in cropping, sorghum does not exceed 50% of the cropping
programme. No more than three successive sorghum crops are grown in the
same paddock. Wheat and sunflower are also grown regularly. Mung bean is
introduced into the crop sequence where there are few surface stones that
interfere with harvesting. Green manure crops or lucerne may be used for
restoration of fertility.

Differences Between the Two Regions
The area around Capella has a longer history of cropping than Banana. It
was opened up for cropping shortly after World War II; the Banana area began
wide-scale cropping only 20 years ago, after dairying was abandoned. This is
apparent in some of the rules to do with fertilizer use e.g. Capella has a much
longer history of fertilizer use and the farmers there have more complex rule
sets about the best way to use fertilizer.
In Banana, the farmers interviewed use several indicators of the need for
nitrogen (N) fertilizer. These include: soil fertility; the history of wheat protein
content from that paddock; and cropping intensity. Indicators of the need for
phosphorus (P) fertilizer include: soil tests; soil type (together with knowledge
of previous soil tests); and plant symptoms ("yield OK").The farmers in


Vol 5, No. 2, 1995








Figure I: Example of a partial decision tree: winter crop planting and N fertilizer in Capella.


(USE NITROGEN; DON'T USE NITROGEN)


IS SOIL WATER > 70%?

E/ -
YE





IS THE FERTILITY HIGH? DON'T PLANT
-YES DON'T PLANT


IS SOIL NITRATE
N HIGH? NO NEED FOR N


CAN I APPLY
.N AT PLANTING? YES
O,, |- APPLY N AT-
NO PLANTING Is THERE A PLANTING
OPPORTUNITY? PLANT CRO
APPLY N BEFORE
PLANTING






AN INTERROGATIVE APPROACH TO SYSTEM DIAGNOSIS INVITATION TO THE DANCE 73

Table 1: The crop sequence rule sets of Bruce and Sheila, both farmers in
Banana.


Bruce (a cropping system):
I just crop. I have cattle on agistment with the proviso that when we get 2-3
inches or more rain, they've got to go. It depends on what season it is. I can't
say what I am going to plant when; it doesn't work. I like rotating my crops.
Wheat will not grow on top of sorghum unless you bung a heap of fertilizer on
it. I tried Caluna peas one year; they worked well (like a season's fertilizer).
You've got to plan for these things a bit so you are ready when it happens.
There's possibilities for doing anything but the seasons they're the problem.


Sheila (a cattle with cropping system):
I grow sorghum and cattle. I grow wheat occasionally if I don't get the
sorghum in on time (maybe every second or third year). I grow sorghum mainly
for cattle. In some paddocks, I grow oats for fodder. I move the fodder crop
around; its never in the same place twice. Now I'm trying to get some legume
into it. I've thrown in some snail medic when planting wheat/oats; that works
well. I have used lucerne (with pigeon grass once); it lasted 4-5 years. I have
used wheat, lucerne and pigeon grass at the same time (once). Ratoon sorghum
means that I can get more acres of forage. This means that more acres can be
planted to grain [sorghum].



Capella have a more comprehensive rule set for fertilizer use than those in
Banana. Additional rules for N use relate to: routine nitrate N soil tests; the
amount of soil water; seasonal climate forecasts; tax; and equipment (the last
two governing timing rather than need). Additional features for P fertilizer use
include: the level of VAM (mycorrhiza incidence influenced by crop history);
and the nature of the crop (broadleaved).

Representation as Decision Trees
We tried to represent the patterned data as decision trees. We were unable
to do this satisfactorily. We went looking for inter-connecting branches, linked
to other event and decision nodes, but did not find them. We encountered
certain specific problems for a tree model:
linear structure not apparent Although the rules clearly defined a
bifurcation or decision point (and a way of handling that decision), the logical
sequence of decisions did not emerge from our interview data. It appeared that
we were able to see the outer twigs of a decision "bush", but the main limbs


VoL 5, No. 2, 1995






Cox, SHULMAN, RIDGE, FALE, AND GARSIDE


were obscured. It may be that a linear sequential representation did underlie
these rule sets, but this was largely tacit and unconscious. The coherence is
provided externally by the way in which the rule sets interact with the sequence
of events to which the grower is exposed and which progressively unfolds
through time. Coherence does not belong to the rule sets in isolation but only in
the context of the situation in which they are applied; it is an emergent
property. Linearity is not, and need not be, a feature recognized by farmers in
their decision-making behaviour.
interconnections between rule sets The rule sets outlining the way in
which individual technological components were managed appeared to
intersect at various points. Decisions about cattle determine what crops are
grown (both the sequence and the use of opportunity cropping e.g. sorghum
versus sunflower, and planting on a reduced soil moisture profile since the crop
can be fed off to the cattle if it fails) and tillage practice (e.g. less concern about
soil compaction if cattle are kept). The rule sets for tillage practice and
opportunity cropping are inter-linked at several points e.g. adoption of zero
tillage can generate more planting opportunities and, perhaps, longer planting
windows; requirements for stubble cover may dictate crop choice.
links to additional rule sets Other technological components were referred
to by farmers in the course of the interviews. In particular, it became apparent
that marketing and seasonal climate prediction were topics engaging their
interest at the moment. Marketing opportunities are increasingly important
following the progressive deregulation of grain marketing in Australia. The
interest in seasonal climate prediction was heightened because of the length
and severity of the drought in Central Queensland that has lasted since 1991.
simultaneous management of multiple indicators of system performance
The criteria by which the performance of the system is judged is not captured in
a single numeraire (cf. Scott et al., 1976), nor in simple biophysical
measurements such as the status of water and nitrogen. Rather, there are many
variables which suggest how the system is performing in both the short- and
long-term, which are monitored to a greater or lesser extent. These include
revenue, bank balance, water and nitrogen (available N in the short-term; total
N in the long-term), but also the status of weeds and pests, and opportunities
for experimenting and learning.
open systems The systems described by farmers, and captured in the rule
sets, emphasise the open nature of agriculture production systems to the
transfer of materials, information and power. One farmer timed his harvesting
operations to coincide with the school holidays with its attendant increase in
the availability of family labour; opportunity cropping is a recognition of, and a
response to, the replenishment of soil water. Outside information impinges on
the system and is progressively incorporated, however patchily. In a situation of
severe and prolonged drought, for the farmer whose equity falls sharply, the
management of the farming system can be dictated by the bank manager.


Journal for Farming Systems Research-Extension






AN INTERROGATIVE APPROACH TO SYSTEM DIAGNOSIS INVITATION TO THE DANCE 75

stability of the rule sets The rule sets appear to be stable over the period of
our interaction with the farmers (about two years). This is a bit surprising in
view of the severity of the drought they were experiencing which might have
been expected to precipitate more erratic behaviour. We found little evidence of
this. There was a greater tendency to consider spring sorghum 'if a planting
opportunity arose, and to plant on a less-than-full profile of soil water, but these
minor modifications were well within the resolution of the current rule sets. We
suspect that the relative stability of the rule sets is not due to conservative
behaviour, nor to a limited ability to modify rule sets, but rather reflects the
contingent nature of the rule sets within the context of a variable environment.
rule sets adaptive to changing circumstances All the farmers we spoke to
were actively experimenting with their system. Some were starting to think
through the best *ay to use inorganic fertilizer in face of a changing perception
of the importance of fertility decline in their soil. Some were adopting
moisture-seeking deep planting with modified equipment to increase the
frequency of planting opportunities. Some were conducting experiments with
legumes and green manure crops in face of concern about increasing
dependence on nitrogen fertilizer. And some were responding to new
information by incorporating it into their rule sets.
importance of interpretation Because the rule sets are contingent on other
things happening, either within the boundaries of the system or impinging from
the outside world, the ways in which these events are interpreted become of
major significance. Some of the discrepancies between the behavior of farmers
defined by the rule sets and the recommendations of professional science have
to do with differences in the interpretation of the implications of different
behaviors. One farmer in Capella regularly tills the soil following a rainfall
event to keep it "soft and fluffy" because his interpretation of his own
observations (model) suggests that fluffy soil absorbs water faster, even when
the advantages of stubble retention for water infiltration have been
demonstrated by professional advisers using an iconic model viz. a rainfall
simulator. From several incidents, it became clear that farmers and researchers
had a different understanding of the way in which plant roots grow and of their
ability to extract soil water deep in the profile.

DISCUSSION

System diagnosis is an important step to take prior to considering an
intervention. It helps to specify the design characteristics of the intervention
and enables us to develop the relationships needed to carry it through. It may be
a mistake, however, to spend too much time in this phase of an investigation at
the expense of beginning to experiment with novel technologies. It is a first cut
at the design process which will be progressively refined as the investigation
proceeds. However, it is an important step where social science can make a


VoL 5, No. 2, 1995






Cox, SHULMAN, RIDGE, FOALE, AND GARSIDE


significant contribution (Dewalt, 1985; Tripp, 1985; Jones and Wallace, 1985;
Maxwell, 1986; Goodell et al., 1990). Sutherland (1987) discusses the
contribution of sociology to FSR. Qualitative research (e.g. Denzin and
Lincoln, 1994) helps to provide the basis for a sympathetic understanding of
what farmers are trying to do in the way they manage their systems.
Before groups such as the Agricultural Production Systems Research Unit in
Australia can develop relevant and useful decision support, they must develop
and maintain regular activities for system diagnosis. Otherwise there is a real
danger that interventions will be only marginally effective because they fail to
address the most important needs of farmers for information and decision
support. But it may also be that performance could be lifted most effectively
and efficiently by more widespread application of rule sets already used by the
most productive farmers.
Our attempt to pattern the decision-making behavior of farmers in Central
Queensland demonstrated the extent to which their espoused models are
apparently non-linear; non-sequential; relatively straightforward (small number
of rules); coherent (exhibiting few contradictions); inter-linked (across
technological components); contingent (on the open nature of agricultural
production systems); stable (in the face of perturbation such as drought); and
dynamic and adaptive to changing conditions (such as long-term fertility
decline) and the generation of new knowledge (whether from their own
resources or from outside). Different rule sets can be distinguished
corresponding with differences in farmer/farming system, region and
technological components.
These characteristics are not captured within a decision tree framework
without imposing a structure that was not explicit in our discussions with
farmers about their rule sets. Decision rules are invoked as needed, rather than
in an ordered sequence defined by a tree. This limitation applied whether we
examined the rule sets asserted by individuals, or (like Gladwin) the aggregated
rule sets of a group. This difference between our experience and Gladwin's may
be due to the contingent (marginal, responsive) nature of Central Queensland
agriculture, but we suspect that it is a general feature of broadacre dryland
farming in our region.
Phillips and Gray (1994) are similarly critical of reductionist
decision/rational action models of farming practice because of their lack of both
cultural consistency and social/spatial specificity. Alternative forms of
knowledge representation and decision rules (e.g. networks, blackboard
systems) may provide a more fruitful approach than linear or tree structures.
Other approaches to this problem include the participatory development and
use of pictorial representations such as resource maps (e.g. Lardon and
Albaladejo, 1990; Lightfoot and Minnick, 1991), and soft systems methodology
(e.g. Macadam et al., 1990; Mills-Packo et al., 1991). The values underlying
both these methodologies also influenced the way in which we approached this


Journalfor Farming Systems Research-Extension






AN INTERROGATIVE APPROACH TO SYSTEM DIAGNOSIS INVITATION TO THE DANCE 77

activity: our concern for gaining understanding of the points of view of the
various participants in a problem situation (the value of multiple perspectives);
an a priori rejection of the notion that any single representation would prove
adequate by itself for any substantive intervention in a human activity system,
even though this may be essential to initiate negotiation about change (the use
of multiple models); and acceptance of the fundamental importance of
communication in bringing about change (through the successive re-negotiation
of meaning).
But the argument is not that we were insufficiently skilled to impose a tree-
like structure on our data. Rather, that a tree is an inadequate metaphor for our
purpose, which is to bring about directed change in the management of
agricultural production systems. The rules that farmers assert about the way in
which they manage various technological components, and our behavior in
engaging with farmers about their rule sets, can be represented as a dance. The
rule sets can be learned (perhaps by treading on toes, even if inadvertently!).
We engage in the dance in order to learn the rules. As we proceed to engage
new partners, we propose rules that we have learnt from previous partners.
These new rules may or may not be the ones used by our new partner and s/he
will tell us if we propose an inappropriate step. We learn the rules by breaking
them within the notional dance. This accords with Garfinkel's
ethnomethodological approach to learning the rules governing the bounds of
acceptable behavior (Garfinkel, 1967). The dancing lessons that we were
involved in are extractive in that we were seeking to learn someone else's rules
for their own dances. But the dances are also inherently participative: as we
become more skilled in the steps, we may begin to influence the conduct of
farmers' own dances.
Another way of learning the steps is deliberately to seek out people whose
behavior, in their own mind, departs from accepted norms. Some farmers are
even more adept than us at breaking the rules in the notional dance. We can
learn about the old steps from their experience in developing new ones. New
steps developed by practitioners are likely to be particularly powerful because
they incorporate local constraints and opportunities whicl outsiders may miss
(e.g. Loevinsohn et al., 1994).
There are many dances going on at once, each with its own rules. Thus,
farmers have different dances governing the use of different technological
components; and different farmers employ different dance steps for each
component. Most interestingly, when farmers were presented with other
farmers' dances, they could name them ('That's what Joe does."). Farmers are
often aware of the rules used in other people's dances: there may be some steps
that can be incorporated into their own routines. They are certainly aware of
how their rule sets depart from the norm for farmers in their area, and will offer
explanations for such discrepancies often in a self-deprecating way. They are


VoL 5, No. 2, 1995






Cox, SHULMAN, RIDGE, FOALE, AND GARSIDE


keen to learn about the rules of other farmers. All this provides additional
evidence that farmers recognize multiple farming systems.
Researchers are involved in co-producing/choreographing/rehearsing/
performing multiple dances with different farmers. The rules can be, and are,
modified as the dance proceeds: it can evolve and adapt to changing
circumstances. In order to change the rules, change agents have to engage their
partners in the dance and discover its rules. We need to demonstrate the new
steps, and what advantage they have over the old steps. The new steps may
require dancing faster or jumping higher, perhaps beyond the dancing skills of
our partners or requiring special apparatus, in which case the new steps may
not be adopted. Sometimes, the new steps will work. They will fit with the
existing repertoire to provide a more elegant and polished performance. Often,
new steps will be adapted before being incorporated by the farmers into other
dances.
In Central Queensland, we did begin to introduce some new steps: use of
moisture seeking planting equipment when a farmer was faced with a marginal
planting opportunity; introduction of a farmer from the Darling Downs who
knew different dances to do with zero tillage; use of deep soil coring. But,
because of the severe drought experienced throughout the study period, we had
limited opportunities to perform emergent steps.
The rule sets, made explicit by the dance, do provide useful models of
farmer decision-making. They help to specify criteria for the design of revised
models through the specification of:
* a template change can be achieved either by presenting the existing rule
set with a new situation (e.g. additional information about soil nitrogen), or
by incrementally modifying the rule set (e.g. so people interpret additional
information differently);
* a baseline by which to judge the marginal improvement claimed for
professional models which need to address more effectively and efficiently
the situation currently managed using a particular rule set e.g. use of a
decision support system;
* a source of insights through demonstration of the possibilities associated
with novel steps through the comparison of different rule sets, including
those of professional researchers e.g. the use of moisture-seeking planting
equipment;
* entry points for change where the current steps fall down, or where the
superiority of alternative steps is evident e.g. the justification for improved
soil monitoring in a situation where soil fertility has declined to a point
where yield and quality may now be responsive to fertilizer use;
* and the basis for substantial dialogue between people who know different
dances e.g. participation in a joint programme of on-farm research.
A knowledge representation based on production rules has several
shortcomings. For other than trivial problems, its behavior is difficult to predict


Journalfor Farming Systems Research-Extension






AN INTERROGATIVE APPROACH TO SYSTEM DIAGNOSIS INVITATION TO THE DANCE 79

because this depends on the conflict resolution strategy used (e.g. the ordering
of the rule set); and it is not particularly convenient for the representation of
randomness (Suppes et al., 1994). Indeed, the logical structures that we have
loosely called "trees" (e.g. Figure 1) are not trees at all as this term is more
rigorously defined in graph theory (Wilson, 1972): because a tree has exactly n-
1 edges where n is the number of vertices (nodes); and because a tree cannot
contain any circuits. Thus, in Figure 1, there are several ways of determining
whether nitrogenous fertilizer should be applied which provide parallel
connections between adjacent vertices (i.e. there are more than n-l edges); and,
because the decision to plant is repeated in successive seasons, this tree-like
segment of the rule set is called recursively within a larger decision framework
(i.e. there are circuits).
Dorward (1991), in an example from Malawi, points out that integrated
decision rules in small-holder farm management provide an analytical
approach which may provide more effective communication in farmer-
extensionist-researcher dialogue and facilitate the development of more
relevant extension messages. He distinguishes two major types of decision:
'task decisions', to do with the allocation of resources to particular tasks; and
'checking decisions', which can interrupt tasks before their completion in face
of changing circumstances. This also implies a structure that is not a tree:
because there are multiple links between vertices (to maintain both task and
checking functions); and because the possibility of revising a decision implies
recursion and hence the formation of circuits.
Winograd and Flores (1986) and Wood and Wood-Harper (1993) stress the
importance of interpretation as the sticking point in system design not
decision-making as a choice between pre-defined alternatives. We have found
several examples from our work of a difference in interpretation of a farm
situation e.g. how surface management affects the way in which water enters
the soil, and how the roots of crops grow in relation to spatially distributed soil
resources such as water and nitrogen. These differences in interpretation, if
they can be successfully negotiated, potentially lead to improvements in system
performance. They do not yet depend on increasing complexity in model
specification either to identify them, or to exploit them. Rather they depend on
rendering some model components more transparent, more open to discussion,
and more available for transformation to fit, or dance with, local circumstances.
In system diagnosis, we are attempting to put ourselves in farmers' shoes,
and to appreciate what they are trying to do, so we can report back to
professional colleagues about what the world is like and how we might best
intervene to change it or at least to facilitate change. But this is very difficult.
At best, one can learn to move, or dance, with others. That requires
establishing a relationship with them; it cannot be done vicariously. The
establishment of relationships directly between participants is needed for the
development of a system of action. We are not advocating that a small number


Vol. 5, No. 2, 1995






Cox, SHULMAN, RIDGE, FOALE, AND GARSIDE


of professional researchers should form regional dance teams in order to learn
to dance with everybody else; this would be cumbersome and costly. But it is
important that we learn to dance, and to dance with a range of different
partners, if our intention is to influence the decision-making behavior of others
in any particular direction and to any great extent.
Janesick (1994) talks about the dance of qualitative research design: 'The
qualitative researcher is like the dancer, then, in seeking to describe, explain,
and make understandable the familiar in a contextual, personal, and passionate
way."
The dance metaphor for communication is functional in that it captures the
combination of stability and fluidity that we found in the rule sets. It
emphasises the requirements for interaction and engagement in order to effect
change. The dance satisfies the three criteria that Casti (1989) proposes as
characteristic of a good model: its simplicity (the dance metaphor is readily
understandable both by researchers and researched); it agrees to a reasonable
degree of accuracy with most of the observed data (it captures in an elegant
way both what farmers do, and how farmers and researchers communicate
about what they do); and explanatory power (it describes what we need to do in
order to effect change in the actions of both partners).
The dance metaphor is consistent with the actor-oriented sociology of Long
and Long (1992). Unlike systems models (e.g. Ruling, 1988) that emphasise
information flow and linkage between different parties in an agricultural
knowledge and information system (or AKIS), the guiding notions of the dance
are discontinuity between actors in their knowledge of the steps (engagement
and dis-engagement in different dances, not linkage), and transformation of
meaning (not the transfer of information). Discontinuity because of the way the
dance metaphor distinguishes and values different ways of doing things,
transformation of meaning through the progress of the dance between partners
with different levels of skill, and preference for different dance styles.
It is a counter to the "central source" perspective of many FSR practitioners
and planners that Biggs and Farrington (1993) argue has been the main
constraint to sustained implementation of FSR. and is compatible with their
political economy perspective. Knowledge emerges and develops, and behavior
changes, as a product of the interaction conceptualised as dance.

CONCLUDING REMARKS

Gladwin's ethnographic decision tree modelling helped to explore the richness
underlying individual decision-making, not captured by an exploratory RRA,
and provided a readily-understandable representation. However, excessive
preoccupation with the tree representation (and particularly its use in a
predictive manner) can impede understanding about the behavior of managed
systems, and restrict the access of change agents to them. Our approach is


Journal for Farming Systems Research-Extension







AN INTERROGATIVE APPROACH TO SYSTEM DIAGNOSIS INVITATION TO THE DANCE 81


consistent with postmodern ethnography (Glesne and Peshkin, 1992; Sarup,
1993) in our concern with how researcher and researched affect each other,
how the research act is socially constructed, and the conscious choice of
metaphor.
The dance metaphor provides a model of communication that captures in an
intriguing way the activities of both data collection and intervention. Indeed,
system diagnosis and subsequent attempts to effect change cannot easily be
separated: diagnosis, treatment and prognosis go hand in hand. It emphasises
the primacy of action, and of action as a source of understanding (Penman,
1994). It captures the need for engagement in order to effect change. It is
simultaneously participative and extractive. Qualitative research will have an
increasing role in system diagnosis, particularly as system diagnosis and the
management of intervention converge within a problem-based methodology
(Robinson, 1993).
The proposed change in metaphor (from a tree to a dance) is an important
step in changing the behavior of farming systems researchers: the way in which
they view the world; the way in which the potential for change in the
management of agricultural production systems is negotiated with other
stakeholders; and how we act together to effect change in the probability that
certain outcomes will be realized.

ACKNOWLEDGMENTS

The Land and Water Resources Research and Development Corporation, the Grains Research and
Development Corporation and the Rural Industries Research and Development Corporation of Australia
all supported financially the work described here. We are grateful to Lisette Ackhurst for technical
support. The paper has benefited from the valuable suggestions and comments of Warwick Easdown,
Duncan Lowes, Neil Macleod, Peter Carberry and Bob McCown.

REFERENCES

Alty, J.L., and M.J. Coombs. 1984. Expert systems. Concepts and examples. Manchester: The
National Computing Centre.
Argyris, C., R. Putnam, and D.M. smith. 1985. Action science. Concepts, methods and skills for
research and intervention. San Francisco: Jossey-Bass.
Beebe, J. 1985. Rapid rural appraisal: the critical first step in farming systems research. FSSP Network
Paper No. 5. Gainsville: University of Florida.
Biggs, S.D., and J. Farrington. 1993. Farming Systems Research andthe ruralpoor: a political
economy approach. Journal for Farming Systems Research-Extension 3 (2):59-82.
Casti, J.L. 1989. Alternate realities. Mathematical models of nature and man. New York, NY:
Wiley-Interscience.
Collinson, M. 1981. A low cost approach to understanding small farmers. Agricultural
Administration 8: 433-450.
Cox, P.G., AL. Garside, and R.L. McCown. 1993a. Participatory design of an on-farm agronomic
research programme in Central Queensland. Proceedings 7th Australian Agronomy Conference,
Adelaide, 19-24 September 1993. pp. 325-327.


VoL 5, No. 2, 1995







Cox, SHULMAN, RIDGE, FOALE, AND GARSIDE


Cox, P.G., A.L. Garside, and R.L.McCown. 1993b. Market research for decision support in dryland
crop production in Central Queensland a preliminary report. Proceedings 7th Australian
Agronomy Conference, Adelaide, 19-24 September 1993. pp. 332-335.
Denzin, N.K., and Y.S. Lincoln. 1994. Handbook of qualitative research. Thousand Oaks, CA:
Sage.
Dewalt, B.R. 1985. Anthropology, sociology and farming systems research. Human Organization 44
(2): 106-114.
Doorman, F. 1991. A framework for the rapid appraisal of factors that influence the adoption and
impact of new agricultural technologies. Human Organization 50 (3): 235-244.
Dorward, A 1991. Integrated decision rules as farm management tools in smallholder agriculture in
Malawi. Journal of agricultural Economics 42 (2): 146-159.
Dunn, A, and I. Gray. 1992. Farmer perceptions of dryland salinity in the southern NSW wheatbelt.
Australian Journal of Soil and Water Conservation 5 (2): 44-49.
Feldstein, H.S., and J. Jiggins (Eds.). 1994. Tools for the field Methodologies handbookfor gender
analysis in agriculture. West Hartford: Kumarian Press.
Garfinkel, H. 1967. Studies in ethnomethodology. Englewood Cliffs, NJ: Prentice-Hall.
Gladwin, C.H. 1976. A view of the Plan Puebla: an application of hierarchical decision models.
American Journal of agricultural Economics 58 (5): 881-887.
Gladwin, C.H. 1983. Contributions of decision-tree methodology to a farming systems program.
Human Organization 42 (2): 146-157.
Gladwin, C.H. 1989. Ethnographic decision tree modelling. Sage University Paper Series on
Qualitative Research Methods, Vol. 19. Beverley Hills, CA: Sage.
Glesne, C., and A Peshkin. 1992. Becoming qualitative researchers. An introduction. White Plains,
NY: Longman.
Goodell, G., K.L. Andrews, and J.I. Lopez 1990. The contributions of agrono-anthropologists to on-
farm research and extension in integratedpest management. Agricultural Systems 32: 321-340.
Goodwin, P., and G. Wright. 1991. Decision analysis for management judgement. Brisbane: Wiley.
Hamilton, N.A. 1991. Attitudes and behaviours offarmers and other stakeholders involved in
southern Queenslandfarming systems. An analysis by modified Rapid Rural Appraisal.
Information Series QI91026. Brisbane: Queensland Department of Primary Industries.
Hildebrand, P.E. 1981. Combining disciplines in rapid appraisal: the sondeo approach. Agricultural
Administration 8: 423-432.
Ison, R.L., and P.R. Ampt. 1992. Rapid Rural Appraisal: a participatory problem formulation method
relevant to Australian agriculture. Agricultural Systems 38: 363-386.
Janesick, V.J. 1994. The dance of qualitative research design. Metaphor, methodolatry, and meaning.
In: Denzin, N.K. and Lincoln, Y.S. (Eds.), Handbook of qualitative research. Ch. 12, pp. 209-
219.
Jones, J.R., and B.J. Wallace (Eds.). 1985. Social sciences and farming systems research:
methodological perspectives on agricultural development. Boulder: Westview.
Kumar, K. 1993. Rapid appraisal methods. Washington: World Bank.
Lardon, S., and C. Albaladejo. 1990. Modeling an agrarian system on a local scale as a tool for farmer
participation in rural development. Journal for Farming Systems Research-Extension 1 (2):
137-155.
Lightfoot, C., and D.R. Minnick. 1991. Farmer-first qualitative methods: farmers' diagrams for
improving methods of experimental design in integrated farming systems. Journal for Farming
Systems Research-Extension 2 (1): 57-69.
Loevinsohn, ME., J. Mugarura, and A Nkusi. 1994. Cooperation and innovation by farmer groups:
scale in the development of Rwandan valley farming systems. Agricultural Systems 46: 141-155.
Long, N., and A. Long. 1992. Battlefields of knowledge. The interlocking of theory and practice in
social research and development. London: Routledge.
Macadam, R., I. Britton, D. Russell, W. Potts, B. Baillie, andA. Shaw. 1990. The use of soft systems
methodology to improve the adoption by Australian cotton farmers of the Siratac computer-based
crop management system. Agricultural Systems 34: 1-14.


Journal for Farming Systems Research-Extension







AN INTERROGATIVE APPROACH TO SYSTEM DIAGNOSIS INVITATION TO THE DANCE 83


Maxwell, S. 1986. The social scientist in farming systems research. Journal of agricultural
Economics 37 (1): 25-35.
McCracken, J.A., J.N. Pretty, and G.R. Conway. 1988. An introduction to Rapid Rural Appraisalfor
agricultural development. London: International Institute for Environment and Development.
Mills-Packo, P.A., K. Wilson, and P. Rotar. 1991. Highlights from the use of the soft systems
methodology to improve agrotecmology transfer in Kona, Hawaii. Agricultural Systems 36: 409-
425.
Murray-Prior, R. 1994. Farmers as scientists: a personal construct theory interpretation of hierarchical
decision models. Paper presented to the 38th Annual Conference of the Australian Agricultural
Economics Society, Victoria University, Wellington, February 7-11 1994.
Penman, R. 1994. The participating researcher. Theoretical implications in practice. Paper presented at
the International Communication Association and Australian & New Zealand Communication
Association Annual Conference, Sydney, July 11-15 1994.
Phillips, E., and I. Gray. 1994. Farming 'practice' astemporally and spatially situated intersections of
biography, culture and social structure. Paper presented at the Annual Conference of the Institute
of Australian Geographers, Magnetic Island, September 1994.
Robinson, V. 1993. Problem-based methodology. Research for the improvement of practice. Oxford:
Pergamon.
Ruling, N. 1988. Extension science. Information systems in agricultural development. Cambridge:
Cambridge University Press.
Sarup, 1993. An introductory guide to post-structuralism andpostmodernism. 2nd Edition.
Hemel Hempstead: Harvester Wheatsheaf.
Scott, M.F.G., J.D. MacArthur, and D.M.G. Newbury. 1976. Project appraisal in practice. London:
Heinemann.
Sutherland, A. 1987. Sociology in farming systems research. Agricultural Administration Unit
Occasional Paper 6. London: Overseas Development Institute.
Tripp, R. 1985. Anthropology and on-farm research. Human Organization 44 (2): 114-124.
van Nieuwkoop, M., T. Defoer, T., and S.S. Hussain. 1994. The contribution of rapid rural appraisals
in the planning of on-farm research and extension activities Agricultural Systems 44: 201-216.
Wilson, R.J. 1972. Introduction to graph theory. New York, NY: Academic Press.
Winograd, T., and F. Flores. 1986. Understanding computers and cognition. A new foundation for
design. Reading: Addison-Wesley.
Wood, J.R.G., and A.T. Wood-Harper. 1993. Information technology in support of individual decision-
making. Journal of nfo Systems 3: 85-101.


VoL 5, No. 2, 1995









SUSTAINABLE PERI-URBAN VEGETABLE
PRODUCTION AND NATURAL RESOURCES
MANAGEMENT IN NEPAL'




Hans G.P. Jansern, DavidJ. Midmore,' and Durga Dutta Poudel


ABSTRACT


A diagnostic survey was conducted by a multidisciplinary
team in Bhaktapur, Dhading, and Bara districts in Nepal to
gain a better qualitative and quantitative understanding of
vegetable cultivation in districts which are important for
Kathmandu's daily vegetable supplies; and to elicit farmers'
perceptions about biotic and abiotic growing constraints. The
survey featured informal, but structured farmer interviews,
and covered farmers' practices and knowledge, levels of
inputs and yields, field observations, and problem
identification. Vegetables in the three survey district are
grown in a large number of highly different cropping
patterns, varying from year-round vegetable cultivation in
Bhaktapur, rice or maize-based systems in Dhading, and
rice-wheat systems in Bara. The use of inputs in vegetable
cultivation is high and at times excessive, particularly for
chemical fertilizers and insecticides. However, returns to
vegetable growing exceed those to rice and wheat manifold,
but risk and capital investment are also considerably higher.
While diseases and (particularly) pests seriously reduce yield,
most vegetable growers suffer from a lack of information

1 All three authors were at the Asian Vegetable Research and Development
Center (AVRDC), Taiwan, R.O.C. at the time this paper was prepared.
2 Research Program on Sustainability in Agriculture (REPOSA-
UAW/CATIE/MAG), Guipiles, Costa Rica.
3 Department of Biology, Central Queensland University, Rockhampton Qld
4702, Australia.
4 University of Georgia, Athens, Georgia, U.S.A.


VoL 5, No. 2, 1995






JANSEN, MIDMORE, AND POUDEL


about pest and disease management, with integrated pest
management (IPM) practices virtually unknown. Similarly,
more and better information is needed regarding appropriate
use of chemical fertilizers. The performance of extension
services in providing the necessary information is inadequate.
Both input delivery and output marketing systems are
inefficient, resulting in scarcities or even lack of inputs at
critical times and high costs for farmers. Particularly, quality
and availability of seed represent a major difficulty for many
vegetable growers. Suggestions are offered for follow-up
research activities as well as government investments and
policy measures.


Background and Justification
Vegetable production in Nepal during 1991/92 is reported as just over 1 million
tons (t) (ASD 1992; VDD 1992; Table 1). The 1992 production is obtained
from an area of 140,500 hectares (ha) (DOH 1992). The latter represents just
over 5% of total arable land in Nepal (Jansen 1992). At approximately 2% of
domestic production (unpublished data from the Agricultural Marketing
Development Division), international trade in vegetables is relatively minor
and mostly with India, partially because of the overvalued Nepali rupee.
However, imports of vegetables exceed exports manifold, with unrecorded
imports from India adding considerably to a negative trade balance in
vegetables of about 3 million US dollars in 1991 (unpublished data from the
Department of Customs). The government has interest in closing the vegetable
trade deficit through increasing commercial production of vegetables near
major foci of demand (so-called peri-urban vegetable production).
Improvements in the management of the natural resources base, better crop
management, identification and introduction of appropriate production
technologies, and better handling and marketing of the products all could
contribute to achieving this goal. Government efforts in these areas center
around the so-called Special Program which focuses on peri-urban vegetable
production. The government tries to make improved seeds, credit, training, and
extension support available to participating farmers.
Average post-harvest weight-based losses in vegetables in Nepal have been
estimated at around 25% (Sharma 1988, Shrestha et al. 1992). Losses for
vegetables produced in the terai and sold in the Kathmandu area are in the
order of 23% for tomatoes, 16% for cabbage, and 12% for cauliflower (Werner
and Subedi 1991). Physical losses occur mainly due to insufficient market
information, poor transportation facilities, and inappropriate handling
procedures.


Journal for Farming Systems Research-Extension






SUSTAINABLE PERI-URBAN VEGETABLE PRODUCTION NEPAL


Table 1. Area, production and yield of vegetables in Nepal, 1991/92

Location Area (103 ha) Production (103 t) Yield (t ha')


Bhaktapur 4.21 30.00 7.12
district

Dhading 2.02 19.42 9.59
district

Bara district 16.40 119.51 7.27

Nepal 140.49 1009.63 7.18


Source: unpublished data from the Vegetable Development Division,
Department of Agriculture, Ministry of Agriculture.


Kathmandu, the major focus of urban demand, has an estimated population
of 1.2 million, which has doubled during the decade of the 1980s (Rekhi et al.
1990), and is now growing at an even higher annual rate of 7.3%, dramatically
exceeding the national population growth of 2.1% per year (national census
1991 and World Bank 1994). Current interest in vegetable production is based
upon the fact that vegetable consumption, in a country where animal forms of
protein and nutrients are minimal, is low -- less than 50 kg per capital per year.
Moreover, vegetable production in Kathmandu Valley is decreasing rapidly due
to increasing urbanization and industrialization. The total volume of vegetables
supplied to markets in Kathmandu Valley has been estimated at only 27,000 mt
(HMG/FAO 1993) which (ignoring seasonality in supplies) translates into an
average yearly per capital supply of less than 25 kg, or less than one-third of the
recommended amount (Ali et al., forthcoming). To ensure vegetable supply
which maintains the acceptable level of around 75 kg per capital per year, while
keeping abreast with predicted population growth, production of vegetables
must be intensified on existing land, especially if new marginal land tracts are
not to be exploited. Such an intensification has been experienced in a number
of Southeast Asian countries in which a demand-led expansion of vegetable
production has taken place.
The high urban demand for vegetables has already stimulated the
emergence of peri-urban vegetable production as a high-input industry. Often


VoL 5, No. 2, 1995






JANSEN, MIDMORE, AND POUDEL


high doses of pesticides and fertilizers are applied that may cause severe
environmental degradation and human health hazards. Commercial vegetable
production areas (Special Program areas) such as the Kathmandu peri-urban
vegetable production are likely to be responsible for a disproportionately large
part of total chemicals used in vegetable production.
Kathmandu receives its supply of vegetables from diverse geographic
regions. During the hot summer months, vegetables are supplied predominantly
from the highlands (i.e., temperate crops, e.g., cabbage), with some
supplementary leafy vegetables from the peri-urban context (including the
"true" peri-urban zone of Thimi/Bhaktapur in Kathmandu Valley where year-
round vegetable production is practiced, and Dhading and Kavrepalanchowk
districts). In addition, some cucurbitaceous vegetables are supplied from the
terai, during the summer. Supplies of vegetables during the winter
(October-February) are dominated by the terai (Nepal and India) with less from
the Valley, when cool temperatures in the highlands preclude production of
most crops.
For the future it is imperative to plan expansion of production in accord
with available national and purchased inputs (especially fertilizer and seed);
with marketing facilities with concomitant inroads into reduction of postharvest
losses; and with development of technology which will ensure equitable
increases in production while ensuring environmental protection (in the widest
sense to include humans too).


METHODOLOGY

Diagnostic methods, combined with RRA-type surveys, have proven to be
extremely useful for eliciting farmers' problems, constraints and opportunities;
identifying potential areas for intervention; and prioritizing preliminary key
researchable issues (Fujisaka 1989, 1991; Tripp and Wooley 1989). A
diagnostic survey lasting one week was carried out in November 1993 by a
multidisciplinary team comprised of personnel from the Asian Vegetable
Research and Development Center (AVRDC) and a number of collaborating
agencies, including the Vegetable Development Division (VDD) and various
other offices of the Department of Agriculture, among vegetable growers in
three major vegetable supply areas for Kathmandu. Members of the team
included an agricultural economist, a geographical information systems (GIS)
specialist, a plant pathologist, an entomologist, a cropping systems specialist,
two horticulturists, as well as one or two local extension specialists in each
survey site. Members of the survey team discussed objectives and methods prior
to beginning the field survey. While working in pairs using informal,
open-ended interviews but guided by a checklist, farmers were interviewed to
understand farmer practices and technical knowledge. The team characterized


Journalfor Farming Systems Research-Extension






SUSTAINABLE PERI-URBAN VEGETABLE PRODUCTION NEPAL


and described the climate, cropping systems, topography, soil, vegetable
marketing channels, agronomic practices, major land-use systems, and
transportation systems of vegetables in the study areas. In addition, the
AVRDC team members visited various government organizations, NGOs,
wholesalers, retailers, as well as national and international research
organizations. Except for the Thimi area in Bhaktapur district where nearly
exclusively year-round vegetable cultivation is practiced, many farmers were
busy harvesting their main rice crop and preparing land for winter vegetables
during the period of the survey. Farmers were mostly interviewed on their
farms or in the villages, and members of the survey teams made sure to visit
farmers' fields. The information gathered was synthesized with some the
(limited) secondary information available.
The survey was conducted among a total of 50 vegetable growers in the
following three districts in Nepal, all of which are important suppliers of
vegetables to Kathmandu city (Rekhi et al. 1990): Bhaktapur, Dhading, and
Bara (Fig. 1). A total of 17 sites were visited based on the importance of
vegetables in prevailing cropping patterns: 5 sites in Bhaktapur (Sano Thimi,
Dhadikot, Shiphadole, Jhoukhel, and Bhakha Bazaar); 7 sites in Dhading
(Simle, Dharke, Khanikhola, Gobre Tar, Gajuri Tar, Baireni, and Benighat);
and 5 sites in Bara (Bishambarpur, Prastoka, Raghunathpur, Amritgunj, and
Hariharpur). At each site, three randomly chosen farmers were interviewed.


OBJECTIVES

The specific objectives of the survey, the first of its kind for vegetables in
Nepal, included the following:
a. To gain an understanding of vegetable cultivation systems in districts which
are important for Kathmandu's daily vegetable supplies.
b. To provide a qualitative and (to a lesser extent) quantitative description of
vegetable growing practices in the three survey districts in terms of cultivation
methods, production systems, yields, input use, and profitability.
c. To elicit farmers' perceptions of major constraints, both those of a
biophysical and of a socioeconomic and institutional nature.
d. To arrive at suggestions for research and government policies aimed at
improving vegetable production.


DESCRIPTION OF THE STUDY AREAS
Location
Bhaktapur district in Kathmandu Valley was chosen primarily because of its
geographical proximity to Kathmandu and because most of it is representative
of the "true" peri-urban vegetable zone. The survey area is a plateau with an


VoL 5, No. 2, 1995






JANSEN, MIDMORE, AND POUDEL


elevation ranging from 1,300 to 1,450 m asl. The road distance from
Kathmandu is less than 15 km for all survey sites in Bhaktapur district.
Vegetable areas in Bhaktapur district do not usually exceed 0.3 ha and may be
as small as 0.05 ha. Average farm size in Bhaktapur is just under 0.5 ha (HMG
1991).
Dhading district is located just outside the valley but with all survey sites
within 1-4 hours' drive (between 25 and 90 km) from Kathmandu. The survey
area is quite hilly, with distinct lowlands and uplands, where elevation ranges
from 450 to 1,000 m asl, i.e., considerably lower than the Kathmandu Valley
plateau. Dhading district is an important supplier of vegetables to Kathmandu
Valley, particularly during the period February-June for tomato, eggplant,
beans, and capsicum (HMG/FAO 1993). At 1.3 ha, average farm size in
Dhading is nearly three times that of Bhaktapur (HMG 1991). Similarly,
vegetable areas (up io 0.5 ha) generally are also considerably larger than in
Kathmandu Valley.
Bara district is located in the terai area, bordering Bihar state in India. The
area is flat, with an elevation of about 100 m asl. Distances from Kathmandu
are substantial at about 300 km, taking 10 hours by road. Bara district is also
an important supplier of vegetables to Kathmandu Valley, particularly during
the period October-February for tomato, eggplant, onion, radish, cauliflower,
cole crops, and okra (HMG/FAO 1993). Farm sizes are larger than in the other
survey districts, with an average of 1.6 ha (HMG 1991). However, due to
relatively high capital requirements for vegetable cultivation and local
extensionists' bias, survey farms were all somewhat larger than the average,
with farm sizes up to 10 ha.

Climate
Temperatures show an inverse relationship to altitude, with mean maximum
temperatures in Dhading significantly exceeding those in Kathmandu Valley,
while in the terai the climate is truly tropical. Summer maximum temperatures
exceeding 30-C and dropping only marginally below 26 C at night during that
period, combined with lots of rain and high relative humidity (>80%), make for
difficulties in vegetable cultivation during this period. However, the terai has a
definite advantage over higher altitude locations elsewhere in Nepal during
some of the coldest months. For example, Kathmandu is heavily dependent on
the terai for winter season fresh tomato (Bhattarai 1992). The climate in
Dhading district is classified as subtropical, with generally lower night
temperatures than in Bara district. Night temperatures in Kathmandu Valley
drop well below 20-C virtually throughout the year, creating larger possibilities
for year-round vegetable cultivation in Bhaktapur district. On the other hand,
night temperatures during the winter months are too low for some fruiting
vegetables (e.g., tomato), creating a dependency during those months on the
terai.


Journal for Farming Systems Research-Extension







SUSTAINABLE PERI-URBAN VEGETABLE PRODUCTION NEPAL


Even though vegetables are mostly grown under irrigated conditions,
rainfall is important, particularly where the irrigation water is mainly supplied
by surface sources with rain water from hills, rivers, streams, and springs
diverted (either directly or through canals) to the vegetable fields. Average
yearly rainfall does not differ dramatically between the three survey districts
(varying from 1,420 mm in Bhaktapur to 1,585 mm in Dhading), although
rainfall variability in Bhaktapur is only about half that in the other two survey
districts. If the number of rainy days gives any indication about rainfall
intensity as well as spread, then rainfall is considerably more evenly spread in
Bhaktapur district as compared to Dhading and (particularly) Bara district.
Relative humidity is somewhat lower in Bara as compared to the other two
survey districts.

Cropping Patterns
The features of the different land types in each of the survey sites in the
three districts are given in Table 2. Besides visual observations, information
regarding soils was obtained from land capability maps generated by a
Canadian-funded land resource mapping project (KESL 1986). More detailed
information can be found in Jansen et al. (1994).


Table 2. Land features in survey sites

District/ Elevation Landscap Soils
survey site m)

Bhaktapur distri

Sano Thimi 1,300 traditional year-round vegetable growing area, dark loamy
flat riverside terraces, borehole irrigation fertile, with
river sand

Dhadikot 1,310 relatively new vegetable area, rice-based systems, dark loamy,
gentle slopes, flat riverside terraces, surface fertile, well
irrigation drained

Shiphadole 1,430 year-round vegetable cultivation, slightly sloping clayey, well
lands, leveled terraces, borehole irrigation drained

Jhoukhel 1,344 rice-based vegetable growing systems, lowlands, dark and
mainly surface irrigated with few boreholes fertile

Bhakha Bazaar 1,330 vegetables mostly grown in rice-based systems, heavy wet
sloping, terraced lands soils
Dhading district

Simla 700 vegetables mostly grown in rice-based systems on alluvial, well
flat riverside leveled terraces, some year-round drained
vegetable cultivation, surface irrigation


VoL 5, No. 2, 1995







JANSEN, MIDMORE, AND POUDEL


Dharke 750 vegetable cultivation in both uplands and in alluvial, well
lowlands along riversides, sloping lands, terraces, drained
surface irrigation but not year-round

Khanikhola 1,000 vegetables cultivated in combined upland maize- clayey,
based and lowland rice-based farming systems, infertile,
well terraced, irrigation not year-round in erosion
uplands prone

Gobre Tar 460 rolling landscape without terracing, vegetables reddish, low
grown in both wheat-maize and rice-based organic
systems, surface irrigation but not year-round matter

Gajuri Tar 450 lowland vegetable cultivation after rice on reddish, low
sloping, well-terraced land with surface irrigation organic
matter

Baireni 450 combined upland and lowland farming systems, reddish, low
vegetables grown in both rice-wheat and maize- organic
based systems, surface irrigation but not year- matter
round in uplands
Benighat 500 vegetables (only winter) cultivated on hill clayey, deep,
plateaus with some leveled flat terraces as well, well drained
in rice-wheat, rice-maize and rice-potato based
systems, good surface irrigation facilities

Bara district

Bishambarpur 91 traditional rice-based vegetable growing area, flat alluvial,
with borehole irrigation fertile, not
well drained

Prastoka 99 flat, close to river with farmer-built temporary alluvial,
dam, some tubewell irrigation as well, vegetables fertile, not
grown in rice-based systems well drained

Raghunathpur/ 90 flat, traditional rice-based vegetable growing alluvial,
Purania/Inaruwa area, borehole irrigation fertile, deep,
not well
drained

Amritgunj/ 81 flat area with vegetable cultivation mostly in alluvial,
Salempur rice-based systems but also some year-round fertile, deep,
canal irrigation for low-lying lands, boreholes for not well
lands higher than canal drained

Hariharpur 82 flat area with vegetable cultivation mostly in alluvial,
rice-based systems but also some year-round fertile, deep
cultivation, canal irrigation for low-lying lands,
boreholes for higher lands


Journal for Farming Systems Research-Extension




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