January 18, 1988
AStrth'esis of AID Exper ien:ce:
Farming Systers Research and Extension (FSR/E) L19 -19
Kerry J. Byrnes (.)
ABSTRACT: The proposed study, by reviewing AID's experience with
farming systems research and extension (FSR/E), will contribute
to the ongoing discussion in the Agency about the potential of
FSR/E, or the the useful elements thereof, to assist AID in
meeting its mandate. The study will identify key factors
influencing the performance of AID-funded FSR/E projects. To the
extent that this may be useful, and where an FSR/E approach is
appropriate, the study will provide information that can be
applied to improve the design, implementation, and evaluation of
FSR/E projects. Where FSR/E projects have not been as successful
as had been expected or desired, the study will identify those
elements of FSR/E which yet would be of value to incorporate into
the design of future AID initiatives in development assistance to
* (' '., -* "0 '- '
Senior Social Science Analyst
Program and Policy Evaluation Division
Center for Development Information and Evaluation
Agency for International Development
SA-18, Room 208
Washington, DC 20523
1. EVALUATION OBJECTIVES AND METHODOLOGY
This report presents a synthesis of United States Agency for
International Development (AID) experience with farming systems
research and extension (FSR/E) projects funded by the Agency from
197 to 198 AID support for FSR/E has been provided through
Centrally funded, non-earmarked support for the Inter-
national Agricultural Research Centers (IARCs) -- an
estimated 15% of IARC budgets supports farming systems
research programs (Anderson, 1985:225);
Centrally funded S&T/Office of Agriculture projects --
e.g., the Collaborative Research Support Projects
(CRSPs), the Farming Systems R&D Methodology Project,
and the Farming Systems Support Project (FSSP);
AID Regional Bureau-funded projects -- e.g., the Africa
Bureau-funded CIMMYT Farming Systems Research Project;
Bilaterally-funded projects -- e.g., AID/Mali-funded
Farming Systems Research and Extension Project.
Although USAID country missions continue to design new
and/or fund ongoing projects having a FSR/E component, AID
funding for the FSSP ended December 31, 1987, thereby terminating
a key mechanism through which the Agency provided support for
FSR/E. With this development, a question arises whether the
current direction and level of AID support for FSR/E is
appropriate relative to the Agency's mandate.
Answering this question is difficult because of the:;
confusion as to what FSR/E is, how FSR/E differs from
conventional approaches to agricultural research and extension,
when FSR/E is appropriate, how to implement FSR/E, and whether
and how to institutionalize FSR/E. As Sands (1986:87) observed,
the ambiguity in terminology and conceptualization of
FSR...has become more acute as the range of activities
encompased by the term FSR has broadened. If the lack of
clarity continues, confusion and misunderstanding about the
objectives, products and role of FSR are likely to discredit
research executed under the name of FSR and jeopadize donor
A second difficulty is the is a lack of information on:
The factors that have influenced the relative success
or failure of donor-supported projects in implementing
The role that FSR/E has played in strengthening the
technology generation and transfer capacity of national
agricultural research and extension systems; and
The impact that FSR/E has had on rural income, food
consumption, and the natural resource base.
A related consideration is the basic issue of what FSR/E can
reasonably be expected to accomplish within a given time frame.
Expectations for FSR/E may have been unrealistic (e.g., someone
oversold the idea). Even if expectations have been realistic,
there is the question of what lapse of time is necessary before
assessing whether FSR/E has succeeded or failed and to what
Finally, where FSR/E projects have been less successful than
had been expected or desired, FSR/E could fall into disrepute in
the Agency, with the attendant risk of the Agency failing to
recognize those elements of the FSR/E approach that are of value
and which should continue to be incorporated into the design of
future development assistance projects in agriculture.
These various-difficulties--the confusion surrounding FSR/E,
the lack of information in the three aforementioned areas, and
the potential discrediting of FSR/E, while failing to recognize
the approach's valuable elements--severely restrict the basis on
which an informed judgement can be made about the direction and
level of support for FSR/E that is appropriate relative to the
Yet the Agency has a vested interest in ensuring that
- experience gained and lessons learned from FSR/E projects are .
available to assist Agency personnel, at the crossroads,.in ':-:..
making decisions about the nature and level of support for FSR/E
that will be in the Agency's best interest. .
The objective of the synthesis is to contribute to the
ongoing discussion within the Agency about FSR/E. More
specifically, the synthesis provides the target audience (defined
below) with information about the performance of past and ongoing
FSR/E projects. This information, in turn, will be useful in
assessing the overall contribution of FSR/E and potential ways in
which FSR/E (or elements thereof) can contribute to the
achievement of AID's mandate. To the extent that this may be
useful, the synthesis also can guide the design, implementation,
and evaluation of FSR/E projects or projects including elements
A second objective of the synthesis is to identify
indicators of FSR/E project performance. Performance may be
defined narrowly in terms.of degree of success in implementing a
project or broadly in terms.of the impact of a project's inputs
and outputs on the project's stated purpose and goal.
1.3 Target Audience
The primary audience which the synthesis is designed to
serve is comprised of AID program and project personnel concerned
with strengthening technology generation and transfer capacity in
national agricultural research and extension systems. Other
relevant audiences include personnel in national agricultural
research and extension systems, agricultural universities and
research institutes in the developing as well as developed world,
private sector organizations (e.g., agricultural cooperatives and
input supply firms), and other organizations concerned with
agricultural research and extension or, more specifically, FSR/E.
This synthesis of AID experience with FSR/E .projects was'
primarily conducted as an exploratory study (Hendricks, 1987).
The actual collection and analysis of data were based on a mix of
literature review, key informant interviews, and a systematic
review (case survey) of existing AID-sponsored evaluations (e.g.,
mid-term evaluations) of completed and ongoing AID-funded FSR/E
projects, including those projects which, although not
specifically called "Farming Systems Research and Extension"
projects, nevertheless included) a significant FSR/E component.
The approach to developing the synthesis was based on a
conceptual model that identifies five cycles in the development
of an AID project. These cycles are: concept, design,
implementation, evaluation, and institutionalization. More
specific questions that could be asked about individual projects
were categorized in terms of one or another of these cycles. A
listing of the more specific questions kept in mind in the course
of reviewing the evaluation documentation is provided in Annex A.
Each of the five identified cycles focuses on a specific
area of concern in AID's overall process of project development
and management, regardless of the project's specific-technical
area. The basic concern underlying each cycle may be stated as a
ConceEt (C) What was the basic technical idea
underlying the project?
-- -Design (D) How was this basic technical idea
translated into a project? (Logical Framework)
I- 1ileentation (I) How was the project managed by the
host-country implementing agency, the technical
assistance team, and USAID?
Evaluation (E) How was the project's performance
measured or assessed?
Institutionalization (I) How did the project provide
for the implementing agency to develop its capacity to
continue to perform the types of activities supported
by the project?
This simple model, referred to herein as the CDIE/I model, as
well as the more specific questions listed in Annex B, provided a
general framework for reviewing a sample of AID-funded FSR/E
projects (or projects including a major FSR/E component)..
Annex B lists the universe of AID-funded projects that were:
identified as FSR/E projects as well as the procedure and -:. ..
criteria followed in selecting the projects to be reviewed.' ;
The data for the study were drawn primarily from the
available evaluation documentation (e.g., special evaluations,
project evaluation summaries, and audit reports) for the projects
reviewed. The basic procedure for reviewing the projects was to
read the documentation, taking notes on content bearing on the
conceptualization, design, implementation, evaluation, or
institutionalization of the project's support for FSR/E.
This information provided the primary data base for
identifying, analyzing, and drawing conclusions concerning the
experience of AID-funded FSR/E projects. While the synthesis was
developed primarily as a desk study, work on the synthesis, from
the "project proposal" through the "project completion" stage was
periodically reviewed by either FSR/E practitioners or Agency
personnel experienced with FSR/E (see Annex C).
1.5 OrgLaization ofReort_
Chapter 1 introduces the study's objectives and methods.
Chapter 2, an overview of the FSR/E concept, provides a vantage
point from which to look back on the AID-funded FSR/E projects
that are reviewed and analyzed, with examples and illustrations
drawn from project experience, in Chapter 3. Chapter 4 sum-
marizes the study's findings, sets forth the major conclusions,
and proposes recommendations for the improved design of AID-
funded projects that include a FSR/E component. Chapter 5
presents "lessons learned," a checklist of criteria to guide
project design, implementation, and evaluation, and a diagnostic
model of factors to consider when designing, implementing, or
evaluating the FSR/E component of AID-funded projects. Chapter 6
briefly discusses potential indicators that would provide
benchmarks for monitoring and evaluating the performance of AID-
funded FSR/E projects.
[Chapters 3-6 in process]
2. An Overview of FSR/E
Some have recommended that the term FSR no longer be used.
....the term FSR may have been used incorrectly or...fallen
into disrepute because of loose usage, but...it is too
important a concept to just abandon. What is important is
to recognize that agricultural research should be geared to
the needs of farmers, and that to do this will require that
research be carried out within a farming systems perspec-
tive. This does not mean that all researchers will be FSR
specialists, nor does it mean that FSR research will be
carried out within a special FSR unit, but it does mean
that...scientists will have a means to focus their work on
the problems that farmers face (Plucknett, et al., 1986:5).
Considerable discussion has surrounded the farming systems
research and extension (FSR/E) concept over the past decade.
However, a consensus on FSR/E is emerging. This chapter presents
a summary overview of the emerging consensus.
The origin of the Farming Systems Research and Extension
(FSR/E) concept lies in pioneering "farming systems" studies
conducted in West Africa, East Africa, and other developing
country sites in the post-Green Revolution era of the 1970s. The
"farming systems" approach gained momentum as the perception grew
that mainstream agricultural research and extension institutions
were following a basically "top-down" approach to technology
development that lacked understanding of the management condi-
tions under which small farmers operate. As a result, technology
development was "guided" by a number of erroneous assumptions, as
follows (adapted from Sands, 1986:88-89):
-- That smallholder farming systems in the tropics and.
sub-tropics are static and primitive. We now recognize
that these are complex, dynamic systems that evolved in
response to particular agro-climatic, ecological, and
That small farmers reject technologies out of sheer
ignorance, traditionalism, or sloth. We now recognize
that small farmers are rational decision-makers; they
often pursue goals and employ criteria for evaluating
technologies distinct from those agricultural
That small farmers seek to maximize yield and profit in
the production and sale of a crop. We now recognize
that sriall-farm households formulate management strate-
gies and make decisions within the context of the whole
economic system exploited by the household, including
cropping, livestock, and off-farm enterprises. Neither
yield maximization nor profit maximization can be
assumed to be the appropriate criteria for assessing
the potential utility and acceptability of a new
technology under the conditions prevailing in
smallholder farming systems.
S That research programs can be effective in generating
broad-based technologies relevant to smallholder
farming systems. We now recognize that many broad-
based technologies were rendered inappropriate by the
great diversity in physical and socioeconomic condi-
tions under which small farmers operate. We further
recognize that if "broad-based" technologies are to be
transferred successfully to small farmers, more
adaptive research is necessary.
In short, and all too frequently, the so-called "improved"
technologies generated in research programs guided by these
assumptions failed to provide the farmer with any incentive to
adopt the subject technologies, given the management conditions
under which he or she operated.
Responding to this situation, a growing number of FSR
practitioners argued: C1) that development of improved
technology for small farmers must be grounded in a knowledge of
the existing farming system; and (2) that technology must be
evaluated not only in terms of technical criteria but also in
terms of the socioeconomic circumstances of the farming system.
FSR programs initiated at various locations during this period
began to provide evidence that multidisciplinary teams comprised
of natural and social scientists could effectively identify
opportunities for appropriate technology change among farmers.
The early work of farming systems pioneers such as Norman,
Collinson, Hildebrand, and others, as well as research programs
initiated by the (IARCs, played a formative role in the origin
and evolution of FSR. Since the "early days" of the farming
systems pioneers, the FSR concept has continued to evolve with
implementation and practical experience. One sign of this was
growing awareness that crop-based approaches to FSR (e.g., rice-
based cropping systems research) risk neglecting important,
interrelated components (e.g., livestock) of a farming system.
Another sign was increasing recognition that the agricultural
productivity of and resource use efficiency in a farming system
should be measured in terms of various limiting constraints
(albeit land, labor, time, or whatever) on the system.
A further sign of the continuing evolution of the FSR
concept was the addition of the "/E" to the earlier, more
narrowly defined concept of FSR as "an approach to research" and
a "normal part of the agricultural research process" (Plucknett,
1987). However, while FSR is certainly not a new science or
discipline, it has been more than simply "an approach to
research" or a "normal part of the agricultural research
process." FSR practitioners have sought not only to conduct
research on and increase knowledge of farming systems but also to
use this knowledge as a basis for bringing about productivity-
and income-increasing change in the farming systems studied.
Viewed in this light, FSR is an approach to and an integral
part of the overall agricultural innovation and technology
management process. For this process to be effective, FSR must
be linked not only with extension CFSR/E) but also with the full
range of agricultural support institutions governing the speed
with which improved technology is generated, tested, evaluated,
adapted,' disseminated, adopted, and diffused in an agricultural
While numerous terms and acronyms' have been used to refer to
the "farming systems" approach, the "FSR/E" acronym is used here
because it explicitly addresses the need for links among farmers,
extension workers, and researchers (Poats, et al., 1986).
2.2 Defining_ ESR/E _and "Farming_Sstem
Farming Systems Research and Extension (FSR/E) is an
approach to the development of "farming systems" that seeks,
through on-farm research and associated extension activities, to
test, adapt, integrate, and disseminate new technologies for
adoption by resource-poor farmers. A "farming system" is:
A unique and reasonably stable arrangement of farming
:- .enterprises that the household manages according to well-
- defined practices in response to physical, biological, and
socioeconomic environments and in accordance with the house-
hold's goals, preferences, and resources. These factors
combine to influence the output and production methods.
More commonality is found within the system than between
systems. 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., 1982:214).
In conducting research on a farm as a system, FSR/E focuses on:
the farm family's attributes (for example, goals,
preferences, skills, access to resources, choice of
productive activities, and management practices);
the interdependencies among system components which
farm family household rermbers are able to control; and
the interaction of these components with the physical,
biological, and socioeconomic factors not under the
household's control (Shaner, et al., 1982:13).
The scope of FSR/E tends to be more limited than that of
integrated rural development (IRD) which focuses on a broadly-
defined set of development problems. FSR/E, in contrast, focuses
on a more narrowly-defined problem--developing improved
agricultural technologies and disseminating these technologies
for adoption by farmers.
FSR/E also may be distinguished from what has been called
the Farming Systems Approach to Infrastructural Support and
Policy (FSIP). Productivity may be improved not only by
developing and disseminating relevant technology (FSR/E) but also
by implementing appropriate policy and support systems (FSIP).
FSR/E is a strategy aimed at developing and disseminating
improved agricultural technologies at the farm level. The
principal product of FSR/E is technology and the primary clients
are limited resource farmers. FSIP operates at a more macro
level than FSR/E and attempts to analyze and influence policy
and/or the progress of institutions which may affect small
farmers. The principal product of FSIP is information, and the
primary clients are policy makers and managers of services and
infrastructure (Hildebrand and Waugh, 1983).
2.3 Goals of FSR/E
Nearly a decade ago, the Technical Advisory Committee (TAC)
to the Consultative Group on International Agricultural Research
(CGIAR) commissioned a Review Team to analyze the farming systems
research (FSR) programs at the International Agricultural
Research Centers (IARC)s. The overall goal of FSR, in the view
of the Review Team, is "to contribute to the improve-ment of
human welfare through sustainable increased agricultural-,
productivity" (Dillon, et al., 1978:17). Adapting the Review
Team's conception of the more specific goals of FSR, the goals of
FSR/E may be stated as follows (adapted from Dillon, et al.,
1978:17; and Plucknett, 1987):
To understand better the problems and needs of farmers,
especially farmers with small amounts of land or land
located in marginal environmrients;
To improve the efficiency of the agricultural research
process by focusing research on the problems and needs
of farmers so to develop improved technology;
To assess the interaction among technologies and
between technologies and the environment, thereby
improving the appropriateness and relevance of new
To ensure that new technologies contribute to the long-
term maintenance and enhancement of agricultural
To facilitate communication among farmers, researchers,
extension (or development) agents, and representatives
of other agricultural support institutions; and
To assist in the formulation of development policies
and methods that ef-fectively address the problems of
2.4 ObJectives of FSR/E
The TAC FSR Review Team proposed that a well-structured FSR
program should aim at meeting a number of objectives that are
also relevant to this paper's more broadly defined concept of
FSR/E. These objectives are (adapted from Dillon, et al., 1978;
Plucknett, et al., 1986; and Plucknett, 1987):
To understand the physical (land including climate) and
socioeconomic environment within which agricultural
production takes place;
To identify and evaluate existing, important.farming
systems in specific physical and socioCconomic environ-
ments, in particular the practice and performance of
these systems; and to improve our understanding of the
farmer's skills, preferences, and aspirations;
-- To improve problem identification (target areas,
constraints, etc.) and opportunities for change in
,~ existing farming systems and thereby to assist in
focusing research on key constraints that limit" '
production, farm income, and their sustainability;
To enhance the capacity of research organizations to
conduct research on priority farming systems' problems
so that they are better able to design improved
To conduct research on potentially improved practices,
principles, system components, or subsystems, and to
evaluate these for possible testing on farms;
To evaluate potentially improved systems, or system
compornentts, on farms in major production areas under
no rmal farm conditions; and
S To assist in extending, monitoring the adoption of, and
assessing the impact and benefits of improved farming
While these objectives imply an active FSR/E program, all
objectives likely would not receive full or equal treatment in a
given FSR/E program.
2.5 Core Characteristics of FSR/E
Basically, FSR/E combines the following characteristics
(adapted from Sands, 1985, 1986; Wiese, 1985; Hildebrand, 1985;
and Fartington and Martin, 1987):
-- FSR/E is farmer-oriented. FSR/E targets small-farm
families as the client group for agricultural research
and technology development. Thus, FSR/E's fundamental
objective is to generate technology relevant to the
management conditions of this client group. This is
done by identifying these conditions before proposing
technological solutions, and by adapting technologies
to local circumstances and needs.
FSR/E involves the client _lou~ as_ articipants in the
research and extension process. FSR/E practitioners
involve and work with client group members (i.e., small
farmers) in the design, implementation, and evaluation
of research and extension activities.
FSR/E recongiizes the locational sfeci ficitYof
technical .and human factors. Client groups are
identified in terms of relatively homogeneous groups of
farming systems in specific agro-climatic zones. These
groupings may be further defined in terms of research,
recommendation, and diffusion domains. The criteria
used to classify farming systems into a domain will
depend on the objective of the FSR/E practitioner. For
example, an IARC may develop generalized categories of
farms grouped largely according to agro-climatic
criteria, while a FSR/E practitioner in a national
agricultural research system, working in a specific
region, may categorize farms according to a set of much
more specific criteria such as product mix, presence of
draft power, and socioeconomic status of the household.
FSR/E i ssstkens-oriented. FSR/E, while viewing the
total farm in a holistic manner as a system of natural
and human components, focuses on a specific production
subsystem in order to evaluate (1) interactions between
that subsystem and other subsystems, and (2) the
potential for and impact on the farm of introducing a
change in the technology of the target subsystem.
FSR/E is E_ a "pblaem-sclving" _aroach. Once a region's
farming systems have been grouped into homogeneous
agro-climatic zones (domains), FSR/E identifies the
limiting technical, biological, and socioeconomic
constraints to improved production. Then technologies
potentially effective in removing or relaxing identi-
fied constraints and feasible for the client group of
farming households to adopt are proposed and tested.
- ESR/E tests technologies in on-farm trials. On-farm
collaboration between farmers and FSR/E practitioners
provides each with a deeper understanding of the
farming system and the farmer's decision-making
criteria, and allows for potentially improved tech-
nology to be evaluated under the environmental and
management conditions in which it will be used.
S FSR/E is interdisciplinary. Collaboration among agri-
cultural and social scientists facilitates identifica-
tion of the conditions under which small farmers
operate; accurate diagnosis of constraints; and design,
conduct, and evaluation of research and extension
activities aimed at developing and introducing improved
technologies suitable to the client group of farmers.
FSR/E _comlp eIen.tsL n:ort replaces_ main st r eam_ commodity
and disciiinary Tagricultural research. FSR/E draws
upon "the body of knowledge" of technologies and
management strategies generated by discipline and
commodity research and adapts this knowledge to
specific agro-climatic environments and socioeconomic
circumstances of a relatively homogeneous target group., :
FSR/E,_La _gdnami c and iterative groLess.2_provides
feedback. for shaiEng_ research priorities and a
cultural policies. FSR/E provides information on how
new technologies perform under farm-level conditions as
well as on the farmer's goals, needs, priorities, and
criteria for evaluating technologies. Results of one
season's trials generate hypotheses for testing in the
next. Further, trial results provide an input to the
setting of on-station research priorities as well as to
the formulation of regional- and national-level policy.
2.6 Stages of FSR/E
FSR/E entails five stages (adapted from Norman and
Collinson, 1985; and Sands, 1986:94-96); (1) diagnosis or
description, (2) design or planning, (3) testing or experimenta-
tion, (4) extension or recommendation and dissemination, and (5)
monitoring and evaluation. In practice, boundaries between
stages tend to overlap because of the dynamic and iterative
nature of FSR/E as an R&D (research and development) process.
2.6.1 Diageqosis or Description
During this stage, the farming systems of a region are
examined in relation to the total environment, the constraints
farmers face, and the potential for change in the systems. While
various-methods of data collection are used, four basic steps are
followed: (1) a review of secondary sources for basic data and
descriptive information on the target region, (2) the identifi-
cation of recommendation domains or target groups of farmers, (3)
an exploratory survey or reconnaissance of the region, and (4) a
formal verification survey.
2.6 .2 Design or Plannin
During this stage, potential strategies are formulated to
deal with the constraints identified in the descriptive or
diagnostic stage. Here the "body of knowledge" of past research
(for example, experiment station trials) as well as farmers'
knowledge play an important role in identifying potential
technologies to deal with the identified constraints. Also
important at this stage is the ex ante evaluation of a technology
or practice with regard to its technical feasibility, economic'
viability, and social acceptability for the target region.;. ,
2.6.3 Testin_ or Exper i mentation
During this stage, technologies identified in the design
stage are tested under farm conditions to identify:
the step-wise modifications...which...will allow farmers to
exploit the available biological resources more efficiently,
and which...are both feasible and attractive for farmers to
adopt.... On-farm experiments test the proposed technolo-
gies and adapt them to local conditions. They...fine-tune
the...technology to farmers' needs and circumstances in a
two to three year experimental process. Early trials are
usually managed by researchers with farmers' cooperation.
As the technology becomes more refined, it is tested and
evaluated in farmer-managed trials (Sands, 1986:95).
The farm family's participation in on-farm trials is
critical'. Farmers evaluate new technologies under their own
management conditions. These evaluations are channelled to the
research station to help scientists formulate more realistic and
relevant research priorities. Concurrently, FSR/E practitioners
gain knowledge and insight on the farming system, farmers'
knowledge of their environment, and farmers' management
strategies and resource allocation priorities and decisions.
2.6.4 Extension or Recommendation and Dissemination
During this stage, adapted technologies are disseminated
through extension to other farmers within the recommendation
domain. Where extension personnel have been actively involved in
the earlier FSR/E stages, they wil.1 know how to use the tech-
nology, the farming systems for which the technology is relevant,
how farmers respond to the technology, and how to introduce the
technology to farmers most effectively.
2.6.5 Mlonitoring_ and Evaluation ,..
During this stage, which occurs throughout the FSR/E
process, the pattern of farmer adoption of technology is
monitored as a check on the technology's relevance and utility.
Within resource limitations, the FSR/E practitioner obtains data
on the technology's impact (for example, impact on the pattern of
demand for labor at the household, community, and regional
levels. Such information as may be gleaned is used as a guide in
setting priorities for future agricultural research as well as
for agricultural policy and other agricultural support
institutions serving small-farm agriculture.
2.7 Data Sources in FSR/E
Data collection and analysis are essential in agricultural
research. However, the specific data sources used in FSR/E can
vary depending on the nature of the data required. Three types
of data sources may be identified: base data studies, research
station studies, and on-farm studies. Collection and analysis of
data in these areas may occur sequentially or cyclically but
generally proceed concurrently with interaction and feedback
(Plucknett, et al., 1986:7).
2.7.1 Base Data Studies
Base Data Studies focus on the collection, collation, and
analysis of data on the factors characterizing the environment
and farming systems of a region. These studies usually rely on
secondary data, complemented by on-site investigation, to
describe and understand existing farming systems. Base data
studies are particularly useful in agroecological zoning,
typological classification of existing farming systems, and
identifying resource constraints and opportunities for
2.7.2 Research Station Studies
Research Station Studies focus on generating new tech-
nologies, designing technological components for new systems, or
modifying existing systems. Although usually carried out at a
research station, the orientation of the research, in comparison
with conventional commodity or discipline research, is toward the
farming system and the interactions within that system rather
than to the commodity or discipline per se.
2.7.3 On-Farm Studies -.. .* -: :..
On-Farm Studies focus on developing research at the farm
level, including on-farm experiments (or trials), studies of
technology adoption, and assessment of the impact of new
technology. Experiments or trials conducted at the farm level
may be fully controlled by the researcher, jointly controlled by
the researcher and the farmer, or fully controlled by the farmer.
2.8 Tyfes_ of FSR/E
While FSR/E initiatives may vary in terms of the specific
data sources or combination of data sources used in research on
farming systems, they also may vary in terms of the specific type
of problem the initiative aims to solve. Thus, the relative
emphasis placed on research or extension varies from one type of
FSR/E to the next. For example, farming system component
research (FSCR as described below) places little (or no) emphasis
on extension. Yet FSCR may be an important step in developing
technology components that are subsequently tested by extension
workers in on-farr trials as a central activity of another type
of FSR/E, namely, farming systems adaptive research (FSAR as
Sands (1986) identified six types of FSR/E: farming systems
analysis, farming systems adaptive research, farming system
component research, farming systems base-line data analysis, new
farming systems development, and farming systems research and
2.8.1 Farming Syst iems Analysis
Farming Systems Analysis (FSA) aims at in-depth, quantita-
tive description of the structure and functioning of existing
farming systems, in order to quantify stocks and flows and
understand the structure of interactions within the system. Key
data sources include On-Farm Studies and Base Data Studies. The
typical product of FSA is a model of the system. FSA is
basically what Simmonds (1985) called Farming Systems Research
2.8.2 FarLming SSstemsAdaptive Research
S-. Farming Systems Adaptive Research (FSAR) aims at increasing
the farming system's productivity through the development of
technology adapted to farmer's circumstances. While FSAR takes
the farming system as the unit of analysis in the descriptive
stage, the design and testing stages more likely focus on a
particular subsystem as a potential point of leverage. Key data
sources in FSAR include On-Farm Studies and Research Station
Studies, supplemented and/or guided by Farming Systems Analysis
and/or Farming System Component Research (described below). On-
Farm Studies (e.g., trials) provide input for the design of
research station studies (e.g., experiments).
FSAR is another name for what Simmonds (1985) and CIMMYT
'Byerlet, et al., 1982; Collinson, 1982) called On-Farm Research
with a Farming Systems Perspective (OFR/FSP). This is the type
of FSR/E most frequently conducted under the name of FSR, partic-
ularly by scientists in national agricultural research systems.
2.8.3 FarminS.stterMiCofLonent Research
Farming System Component Research (FSCR) refers to station-
based, applied and adaptive research on farm subsystems or
components designed to support Farming Systems Adaptive Research
(FSAR). Compared with FSAR's focus on the farming system, FSCR
focuses on a specific subsystem or the management of a specific
resource, with the unit of analysis being the field or plot, not
the farming system. Examples of FSCR would include research on
cropping patterns typical of small farm systems such as inter-
cropping, mixed cropping, or relay cropping; crop-animal inter-
actions; or stable-yielding varieties requiring minimal inputs.
FSCR's research agenda is defined either by a station-based
scientist's diagnosis of a constraint affecting the majority of
farmers in a region or by feedback from a FSAR program. Data
generated by FSAR on the management conditions of farming systems
in a region are used by station-based scientists in isolating
specific problems for more in-depth research and in establishing
more relevant research priorities. The product of FSCR is
prototype technology which becomes part of the "body of
knowledge" upon which FSAR can draw.
Many farming systems research initiatives of the IARCs may
be classified as FSCR. CIAT's Bean Program provides a good
example. The typical Latin American small farmer's practice of
intercropping maize and climbing beans is taken as a parameter in
on-station experiments aimed at selecting improved bean
Another example is the rice-based Cropping Systems Program
of IRRI and the Asian Cropping Systems Network. This program -.
combines FSCR and FSAR in a process called Cropping Systems
Research. Having identified land scarcity as the major
constraint limiting rice production in south and south-east Asia,
the Cropping Systems Program focused on developing technologies
to increase cropping intensity. Component technologies (short-
duration rice varieties and planting techniques that permit
double or relay cropping) generated through FSCR are then tested
by national research systems in FSAR programs aimed at fine-
tuning the technologies to the specific environment and
circumstances of a target group of farming systems.
2.8.4 Farming_ Systes Base-Line Dataa analysis
Farming Systems Base-Line Data Analysis (FSBDA) aims at
developing a classification of major types of farming systems in
an agro-climatic zone and diagnosing the major constraints in
those systems. The objective is to learn as much as possible
about the resources of a region (zone) and to determine how
variation in climatic factors and resources affect agricultural
production. Socioeconomic factors (e.g., population density,
land tenure, etc.) may also be analyzed. Key data sources
include Base Data Studies and large-scale surveys.
Typical FSBDA products are physical resource, climate, and
land use maps useful in classifying the major types of farming
systems in a region. The information may be used by agricultural
scientists to tailor technology development more closely to the
management conditions of a region's farming systems, and by
planners to set general research priorities and to select sites
for more focused FSCR and FSAR.
FSBDA is an in-depth version of the diagnostic or descrip-
tive stage of FSR/E. However, FSBDA (which focuses on an agro-
climatic zone) is executed on a larger scale than FSAR (which
focuses on the farming systems within an agro-climatic zone).
The focus of analysis is on the environment and the general
configuration of farming systems rather than on the internal
organization of a specific type of farming systems. Greater
emphasis is placed on biological and physical rather than socio-
IARCs having regional mandates (for example, ICRISAT, ILCA,
and IITA) have used FSBDA extensively.
2.8.5 New Farmi ngSyst es _Develfopment
New Farming Systems Development (NFSD) aims to generate a
broad-based technology designed to overcome major constraints in
a large agro-climatic zone. In contrast to FSAR (which seeks -to
develop technology suitable for stepwise modification of existing
farming systems), NFSD seeks to bring about revolutionary change
in the farming systems of a region. Farming systems are defined
primarily in physical and biological terms, with socioeconomic
factors largely being left out of the technology design process.
It is assumed that socioeconomic circumstances will have to be
subsequently adapted, most likely through government interven-
tion. Research Station Studies (e.g., on-station experiments)
provide the key data source, although Farming Systems Analysis,
Farming System Component Research, and/or Farming Systems Base-
Line Data Analysis may provide supplementary data.
IITA's program to develop a more stable and productive
agricultural system to replace shifting cultivation in the humid
and sub-humid tropics provides a good example of NFSD. This
research, involving minimal on-farm research, is primarily
station-based strategic and applied component research.
ICRISAT's program to develop watershed management units for
the semi-arid tropics is a second example of NFSD. Technologies
have been developed that improve drainage and enable double
cropping on deep Vertisol soils. While the technology has
produced good results in on-station trials and potentially has
widespread application, major farmer acceptability -problems
emerged in on-farm trials.
This development is not -surprising given NFSD's lack of
attention to socioeconomic factors during the technology design
stage. While the research program defined the watershed
management units in physical and biological terms, establishment
of these units requires that dispersed, individually-owned
landholdings be consolidated into a single resource management
unit. However, the feasibility of such a radical socioeconomic
reorganization within the farming community was not considered
during the technology design stage. Social scientists only
became actively involved in the research at the on-farm testing
stage. Design and development of the watershed management units
could have been facilitated and resources probably used more
effectively if socioeconomic factors and farmers' perceptions of
their needs had been incorporated into the research from the
2.8.6 Farmienq Sgtems Research aAnd Agrlicultural Development
Farming Systems Research and Agricultural Development
(FSRAD) aims to implement farming systems research as an integral
component of a long-term agricultural development strategy and
program for a target region. Although the farming system '(with
its own physical, biological or socioeconomic interactions) is,
the primary unit of analysis, the system's links with the social,
economic, and political environment also are scrutinized to
identify potential leverage points for improved productivity.
Thus, FSRAD includes technological development for major farming
systems as well as reform of agricultural support institutions in
the region. The approach combines research (including mainstream
agricultural research, FSCR, FSAR, and sometimes NFSD) and
development (or modification) of agricultural support institu-
tions, with the objective of increasing overall agricultural
productivity in the region.
In short, FSRAD addresses the common problem encountered in
agricultural development, namely, that a technology, while
technically improved, can be rendered useless because of the lack
of adequately developed agricultural support institutions.
Rather than treating such institutions as given or fixed, as is
usually done in FSAR, FSRAD treats them as variables. Examples
of FSRAD include the Puebla Project in Mexico, the Caqueza
Project in Colombia, and the so-called Francophone approach to
FSR in Africa. FSRAD would appear to be the same as the so-
called Farming Systems Approach to Infrastructural Support and
Policy (FSIP) mentioned in section 2.2.
2.9 Research Strateq yin FSR/E
In organizing and conducting agricultural research, two
basic strategies may be followed: (1) the commodity strategy;
and (2),the resource base strategy.
2.9.1 Commodit Styrataegy
The commodity strategy, which focuses on a particular crop
or livestock enterprise, identifies the steps required to
maintain or, more generally, expand production of the commodity.
Research parameters are defined by the commodity of interest, the
resources used in producing the commodity, and so on. Relevance
of the commodity being studied, as well as the availability of a
suitable resource base, are assumed. The commodity approach
generally is useful in working on better-endowed lands.
2.9.2 Resource Base Strategy
The resource base strategy, which focuses on a particular
resource base, identifies the production system, including its
commodity components, which can best use the available resource
base as it currently is or might be modified. In short, the
resource endowment determines the direction of the .research;
commodity selection, singly or in combination, is determined by
the strengths and weaknesses of the resource base. The resource
base approach is generally useful in working on marginal lands
and depends heavily on integrating the ideas of practitioners
from several disciplines whose practitioners may not be used to
While both strategies can contribute to FSR/E, the greatest
challenge to FSR/E--and perhaps its greatest potential benefits
for resource-poor farmers--lies in the resource base approach.
Developing countries have specific conditions that are unlikely
to be accomodated adequately, if at all, by a conventional
commodity approach and that imply the necessity for employing a
resource base strategy in FSR/E. These conditions include:
Continuing need for increased food supplies created by
increases in population, particularly in tropical and
Increasing use of marginal lands having constraints
that limit the use of technology developed for better
conditions (for example, HYVs bred for optimum
Threat of accelerated, irreversible environmental
Degradation through misuse of marginal lands;
Inability of most farmers, lacking the required power
or means, to identify and communicate their needs to
research agencies and policy makers;
Wide gaps between results achieved on research stations
and obtained by farmers adopting new technologies; and
Increased awareness that women contribute to the food
supply and that agricultural research has seldom
considered women's roles.
On-Farm Studies, FSAR, and FSCR are important ingredients in
a resource base strategy in FSR/E. However, Base Data Studies
and FSA also need to be used for characterization of the
environment, while Research Station Studies and NFSD are needed
in designing new farming systems adapted to the limitations and
problems of marginal lands (adapted from Plucknett, et al.,
1986: 12). .
2.10 Organization of FSR/E .. :
[Alternative models for organizing FSR/E to be summarized here.]
2.11 Im2act and Benefits of FSR/E
The confusion that has surrounded the FSR/E concept over the
years has not made the task of assessing FSR/E's impact and
benefits any easier. Yet, Anderson (1985:226) notes:
FSR workers, if they indeed practice3 what they preach, are
never far from assessing their impact. Whether it is in the
early diagnostic phase of identifying problems, later stages
of testing changes or endloop stages of measuring the
exploitation of modified farming techniques, the close
association with the human elements of FS [Farming Systems]
provides, in principle, a continuous harvest of impact
Impact and benefit assessment of FSR/E ideally takes into
account the extent to which FSR/E-produced farming systems:
-- better achieve the goals of farm families; and
are socially desirable in terms of such criteria as the
sustainability, effects on landless laborers, etc.
There are a number of conceptual problems involved in properly
assessing the impact of FSR/E in terms of these two dimensions.
Anderson (1985), who reviewed these problems, concluded that the
feasibility of either ex post or ex ante assessment is impeded,
if not precluded, by too many conceptual and data problems.
Assessment of FSR/E impact and benefits is only possible in terms
of simple criteria such as:
the- speed and extent of adoption of recommended changes
by farmers; and
-- intuitive assessments of social desirability, guided
where possible by empirical data on such effects as
extent of soil loss, employment levels, and so on.
However, potential benefits are likely including those associated
with better soil and water conservation and use.
Another factor in assessing FSR/E is the role of assistance
mechanisms, ranging from IARCs to individual bilateral assistance
projects, in helping national agricultural research and extension
systems to appreciate and develop capacity to carry our FSR/E
initiatives. Here a key measure is the extent to which FSR/E has
been institutionalized in these systems. The extent to which
FSR/E has been institutionalized in these systems will play an
important role in determinining how quickly and effectively FSCR
and FSAR will be able to transform innovations in biotechnology
into agricultural technology adapted to farming systems.
The following lists potential benefits of FSR/E (adapted
from Plucknett, et al., 1986:18-19; and Anderson, 1985).
FSR/E has strong implications for the organization and
management of agricultural research and extension;
procedures need to be developed that facilitate joint
work without stunting individual initiative and effort.
Given the tendency to concentrate resources on farmers
in resource rich as compared with resource poor areas.
FSR/E can provide a mechanism for ensuring adequate
focus on the latter areas.
FSR/E has the capacity to improve the productivity of
other activities other than those strictly involved in
agricultural production. (For example, a crop becomes
a food only after it has been properly harvested,
stored, cleaned, prepared and, in most cases, cooked.
Improvements in these phases of the overall 'farm-level
commodity processing system might exceed returns from
efforts directed to improving crop yields.)
ETo be revised)
2.12 Lookinqa to the Future of an Evolving_[Con:egt
AID continues directly or indirectly support FSR/E, although
not necessarily in the form of funding projects explicitly titled
"Farming Systems Research and Extension." Key elements of FSR/E
(e.g., on-farm research) are now almost routinely being designed
into AID-funded projects aimed at strengthening agricultural
research and extension programs in the developing countries.
Beyond AID, other donors continue to support FSR/E-type
projects. "IDRC has been very strongly committed to FSR for the
past 15 years....it will stay committed for the next 50." (state-
ment by Andrew Ker, cited in Poats, et al., 1986:76). While the
World Bank has supported the Training & Visit System (T&V System)
as an extension model in many countries, the Bank has begun in
recent years to take a greater interest in FSR (Simmonds, 1985).
One may anticipate that future Bank experience with FSR, building
on T&V-System experience, will lead to additional refinement in
and improved practical application of the FSR/E concept.
Further evidence of the continually evolving nature of the
FSR/E concept may be seen in the emerging emphasis on the role of
resource-poor farmers (RPF) and farmer participatory research
(FPR) in the agricultural innovation and technology management
process. As Farrington and Martin (1987:1) have observed:
...there has emerged a growing concern to understand the
diverse and complex environments in which RPF operate so
that...technology can be tailored to suit their circum-
stances and, more recently, so that farmers' indigenous
technical knowledge (ITK) can be fed into technology
development. It is from these areas of concern..-that the
concept of farmers' direct participation in research (FPR)j
has arisen. :;
The seeds for the emerging emphasis on farmer participatory
research (FPR) were planted in a number of studies. For example,
a potentially important variable in implementing FSR/E is the
nature of farmer participation in "on-farm" activities. In a
study of farmer participation in on-farm testing of new phosphate
fertilization technologies in Colombia, Ashby (1984) found
differences in research outcome depending on the nature of the
farmer's participatory role (nominal vs. consultative vs.
Scientists working in an FSR/E-type mode have formulated
what are, in effect, FPR models. Harwood (1979:33) proposed a
method of small farm development in which "the major emphasis is
on production research, planned and carried out by and with the
farmers on their own fields." Another example is provided by the
"farmer-back-to-farmer" (FBTF) model developed at the Interna-
tional Potato Center (CIP) (Rhoades and Booth, 1982). A third
example is the "farmer-first-and-last" (FFL) model proposed by
Chambers and Jiggins (1986). Common to all of these models is
the recognition of the need to orient research to the farmer as
the client, hence the term "on-farm client-oriented research" in
a recent study of national agricultural research systems
conducted by the International Service for National Agricultural
Research (ISNAR) (CGIAR, 1987:42). Thus, the emphasis on farmer
participatory research (FPR) and on-farm client-oriented research
suggests a variant of FSR/E which may be called "client-oriented
FSR/E" (CO FSR/E).
The emergence of the concept of "on-farm client-oriented
research" is an important step in the evolution of the FSR/E
concept. Indeed, this steps opens the door to finding new ways
to direct, manage, and fund agricultural research and extension.
In several countries, AID is exploring ways to cultivate not only
greater private-sector participation in but also private-sector
support and management of agricultural research and technology
transfer. For example, in Honduras, AID is assisting the
Honduran Agricultural Research Foundation (FHIA). FHIA is a
private-sector organization that conducts research aimed at
developing Honduras' potential to compete in non-traditional
agricultural export markets. In the Dominican Republic, AID is
assisting the newly-created Agricultural Development Foundation
to build its endowment, the income from which will be used to
fund agricultural research on non-traditional agricultural export
The growing emphasis on stimulating private-sector partici-
pation in agricultural research and technology transfer for non-
traditional agricultural export crops helps to bring into relief
that FSR/E could play a more active role in assisting farmers to
identify market opportunities that provide incentives for farmers
to grow new non-traditional market and export crops. A greater
level of activity of FSR/E practitioners in helping farmers to
develop their ability to produce a range of marketable crops
implies another front along which the FSR/E concept could evolve,
namriely, in the direction of what may be termed "market-driven
FSR/E" (MD FSR/E).
The emergence of the concepts of "on-farm client-oriented
research" and "market-driven FSR/E" will create a dynamic that
further shapes the definition of FSR/E as an evolving concept.
Indeed, resource limitations and efficiency considerations will
likely create pressure to find ways, across heterogeneous agro-
climatic zones, to more effectively involve homogeneous groups of
resource-poor farmers in designing, implementing and evaluating
FSR/E. As FSR/E practitioners gain experience working with
farmers and farmer groups, there will be increased pressure and
opportunity for farmer groups or organizations to assume greater
responsibility for designing, implementing, and evaluating FSR/E
in particular and agricultural research and extension in general.
In this respect, innovative approaches will likely be
explored, especially where progress has been or could be made by
providing assistance to strengthen private-sector farmer groups
and organizations (e.g., AID/Bolivia's Private Agricultural
Producer Organization Project). There would appear to be great
potential for farmer organizations to play a more active role in
funding, designing, managing, participating in, and reaping the
benefits of agricultural research, particularly where such
research is carried out in a FSR/E mode. Where farmer organi-
zations begin to play a more active role in agricultural research
and extension, not only participating in agricultural research
(i.e., FPR) but also in setting the direction and priorities of
such research, one may envision that the FSR/E concept will
evolve in the direction of what may be termed "client-directed
FSR/E" (CD FSR/E).
2. 13 Sum'fiary
This chapter has provided an overview of key concepts that
define that field of farming systems research and extension
(FSR/E). Table 1 provides a summary listing of these key
concepts. '' '
.... + ? .: ,,.: .. ".+-..- +: -, ,. ', ... .
Table 1. A Summary of Key Concepts in Farming Systems Research
Core Characteristics of FSR/E
4: Involves client group as participant in research and extension
t Recognizes locational specificity of technical and human
: A "problem-solving" approach
* Tests technologies in on-farm trials
* Complements, not replaces, mainstream commodity and
disciplinary agricultural -research
* Provides feedback for shaping research opportunities and
States of FSR/E
Diagnosis or Description
Design or Planning
Testing or Experimentation
Extension or Recommendation and Dissemination
Monitoring and Evaluation
Data Sources in FSR/E
Base data studies
Research station studies
Farming systems analysis (FSA)
(farming systems research sensu strict)
Farming systems adaptive research (FSAR)
Con-farm research with a farming systems perspective (OFR/FSP)3
Farming systems component research (FSCR) ':
Farming systems base-line data analysis (FSBDA)) ..;...
New farming systems development (NFSD). -
Farming systems research and agricultural development (FSRAD)
Similar to the so-called Francophone approach to FSR in Africa
or the farming systems approach to infrastructural support and
Research Strategy in FSR/E
Resource base strategy
Table 1. Continued
Organization of FSR/E
[alternative models of organizing FSR/E will be listed here]
lrfiacts and Benefits of FSR/E
FSR/E as an Evolving Concept
Client-Oriented FSR/E (CO FSR/E)
Market-Driven FSR/E (MD FSR/E)
Client-Directed FSR/E (CD FSR/E)
Anderson, Jock R.
1985 "Assessing the Impact of Farming Systems Research:
Framework and Problems," Agricultural Administration,
1984 Participation of Farmers in Technology Assessment.
Muscle Shoals, Alabama: International Fertilizer
Byerlee, Derek, Larry Harrington, and Donald L. Winkelmann
1982 "Farming Systems Research: Issues in Research Strategy
and Technology Design," American Journal of Agriucltural
1987 CGIAR 1986-1987 Annual Report. CGIAR Secretariat, 1818
H Street, N.W., Washington, D.C. 20433.
Chambers, Robert and Janice Jiggins
1986 Agricultural Research for Resource Poor Farmers: A
Parsimonious Paradigm. Discussion Paper 220.
University of Sussex: Institute of Development Studies.
Collinson, Michael P.
1982 Farming Systems Research in Eastern Africa: The
Experience of CIMMYT and Some National Agricultural
Research Services, 1976-81. International Development
Paper No. 3. East Lansing, Michigan: Michigan State
Conklin, Frank S.
1982 "FSR/E as a Field Methodology in Third World Countries:
Its Historical Origins, Current Functions, and
Suggestions for Improvement," prepared for the Farming
Systems Research Symposium, Kansas State University,
November 21-23, 1982., -
Dillon, John L., Donald L. Plucknett, and Guy J. Vallaeys.
1978 Farming Systems Research at the International
Agricultural Research Centers. A report to the
Technical Advisory Committee, Consultative Group on
International Agricultural Research. Washington,
Farming Systems Support Project
1987 Farming Systems Research Project Directory (Draft).
Farrington, John and Adrienne Martin
1987 Farmer Participatory Research: A Review of Concepts and
Practices. Discussion Paper 19. Agricultural Admini-
stration (Research and Extension) Network, Agricultural
Administration Unit, Overseas Development Institute.
Small Farm Development: Understanding and Improving
Farming Systems in the Humid Tropics. Boulder, Colorado:
1987 Learning from Development Experience: An Approach to
Synthesizing AID's Existing Information. Center for
Development Information and Evaluation, Agency for
International Development, Washington, D.C.
Hildebrand, Peter E.
1985 "On-Farm Research: Organized Community Adaptation,
Learning and Diffusion for Efficient Agricultural
Technology Innovation," Farming Systems Support Project
Hildebrand, Peter E. and Robert K. Waugh
1983 "Farming Systems Research and Development," Farming
Systems Support Project Newsletter, 1(1)(Spring):4-5.
"Rationalising Mixed Cropping under Indigenous
Conditions: The Example of Northern Nigeria," Journal
of Development Studies, 11:3-21.
Plucknett, Donald L.
1987 "Farming Systems Research at the International
Agricultural Research Centers," presented at the Mid-
Term Meeting of the CGIAR, Montpellier, France, May 19,
Plucknett, Donald L., John L. Dillon, and Guy J Vallaeys..
1986 "Review of Concepts of Farming Systems Research -`- The
What, Why and How," paper prepared for the IARC Workshop 4
on Farming Systems Research, February 17-21,'1986,
Poats, Susan, Daniel Galt, Chris Andrew, Lisette Walecka, Peter
Hildebrand, and Kenneth McDermott
1986 "Farming Systems Research and Extension: Status and
Potential in Low-Resource Agriculture," draft report
submitted to the Office of Technology Assessment,
Congress of the United States.
R.E., and R.H. Booth
"Farmer-Back-To-Farmer: A Model for Generating
Acceptable Agricultural Technology," Agricultural
Sands, Deborah Merrill
1985 A Review of Farming Systems Research, a paper prepared
for the Technical Advisory Committee, CGIAR.
1986 "Farming Systems Research: Clarification of Terms and
Concepts," Experimental Agriculture, 22:87-104
W.W., P.F. Philipp, and W.R. Schmehl
Farming Systems Research and Development: Guidelines-
for'Developing Countries. Boulder, Colorado: Westview
Sirmmonds, Norman W.
1985 Farming Systems Research: A Review. World Bank
Technical Paper Number 43. Washington, D.C.: World
1985 Farming Systems Research: Issues for Evaluation,
prepared for PPC/CDIE, U.S. Agency for International
Development, Washington, D.C. 20523.
i... i i
.-r .- .~-r--
"~' :; '"
::~a-... : -