University of Arkansas to host...
 How FSSP can assist in developing...
 Some common sense about farmer...
 Development oriented research in...
 On-farm research in Morogoro district...
 Impact of a national farming systems...
 Country orientation notebooks for...
 Report of a workshop on non-traditional...
 Building ORF into existing agricultural...

Title: Farming Systems Support Project newsletter
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00071908/00015
 Material Information
Title: Farming Systems Support Project newsletter
Alternate Title: FSSP newsletter
Physical Description: v. : ill. ; 28 cm.
Language: English
Creator: Farming Systems Support Project
University of Florida -- Institute of Food and Agricultural Sciences
Publisher: The Project
Place of Publication: Gainesville Fla
Publication Date: 1983-
Frequency: quarterly
Subject: Agriculture -- Periodicals -- Developing countries   ( lcsh )
Agriculture -- International cooperation -- Periodicals   ( lcsh )
Genre: government publication (state, provincial, terriorial, dependent)   ( marcgt )
periodical   ( marcgt )
Dates or Sequential Designation: Vol. 1, no. 1 (spring 1983)-
Issuing Body: Issued by: Farming Systems Support Project, which is administered by: Institute of Food and Agricultural Sciences, University of Florida.
General Note: Title from caption.
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: UF00071908
Volume ID: VID00015
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 - 10387162
lccn - sn 84011294

Table of Contents
    University of Arkansas to host the 1987 farming systems symposium
        Page 1
    How FSSP can assist in developing a national training strategy
        Page 2
        Page 3
    Some common sense about farmer recommendations and extension advice
        Page 4
        Page 5
        Page 6
    Development oriented research in agriculture
        Page 7
    On-farm research in Morogoro district Tanzania
        Page 8
        Page 9
        Page 10
        Page 11
    Impact of a national farming systems program: Guatemala's institute of agriculture science and technology I.C.T.A.
        Page 12
        Page 13
    Country orientation notebooks for selected countries in Africa
        Page 14
    Report of a workshop on non-traditional experimental designs for on-farm systems experiments
        Page 15
        Page 16
        Page 17
        Page 18
    Building ORF into existing agricultural research and extension institutions
        Page 19
        Page 20
Full Text

Number Four
Quarter. 1986

Farming Systems Support Project Newsletter

University of Arkansas

to Host the 1987

Farming Systems Symposium
The annual Farming Systems Symposium moves to
a new location for 1987, to the Center for Continuing
Education of the University of Arkansas in Fayette-
ville, Arkansas. The University of Arkansas will be
hosting the Symposium in collaboration with Winrock
International Institute for Agricultural Development.
Dates for the Symposium have been set for October
18-21, 1987. The theme for 1987 is "How Systems
Work." Five sub-themes have been identified for
papers and poster sessions: 1) information and com-
munication systems; 2) macro-systems; 3) agroforestry
systems; 4) crop/livestock systems; and 5) crop sys-
Opportunities will be provided for papers and
poster sessions describing FSR/E projects and pro-
grams, and for concentrated preand post-symposium
training on selected topics. Committees have been
formed to develop each subtheme and to establish
guidelines for review of paper and poster sessions
submissions. A call for papers will be issued early in
1987 and the announcement will appear in the next
FSSP Newsletter (Vol. 5, No. 1, First Quarter, 1987).
Inquiries, suggestions for additional or alternative
sub-themes, identification of training needs and
general correspondence relating to the 1987 Sympo-
sium should be directed to:
FSR/E Symposium
P.O. Box 2100
University of Arkansas
Fayettesville, AR 72701

Kansas State University

Receives Recognition
After six years of outstanding leadership in establish-
ing and hosting the annual Farming Systems Sympos-
ium, Kansas State University has stepped down from

this role, pledging its continued support and participa-
tion in future meetings. An award was presented to
Kansas State University, recognizing its contribution
to the advancement of the FSR/E approach through
the Symposium, and for the University's overall
contribution to the Farming Systems Support Project.
This recognition included the presentation of a plaque
with the following inscription:

Presented to Kansas State University
by the Farming Systems Support Project
October 7, 1986

Support Entities:
University of Florida, LEAD
Colorado State University
Cornell University
Iowa State University
Kansas State University
Lincoln University
Michigan State University
North Carolina State University
Pennsylvania State University
Southern Illinois University
Tuskegge Institute
University of Arizona
University of Arkansas
University of Hawaii
University of Illinois

University of Kentucky
University of Minnesota
University of Missouri
Virginia Polytechnic Institute
and State University
Virginia State University
Washington State University
Agricultural Development
Consultants, Inc.
Development Alternatives, Inc.
International Agricultural
Development Service
Research Triangle Institute
Winrock International


How FSSP Can Assist in

Developing a National Training Strategy

The Farming Systems Support
Project has concentrated on training
materials development and course
delivery at the regional level. FSSP
also engages in training activities
at a national level provided funds
for such an activity are available
from the national program or
through other sources such as
USAID or other development as-
sistance agencies.
FSSP experiences provide good
indications of how a national pro-
gram might develop a training
strategy designed to introduce the
FSR/E concept and train a critical
mass of practitioners in order to
initiate FSR/E activities. Such a
strategy is envisioned as having
four stages, each of which is briefly
described below.
1. Orientation
During this stage the national pro-
gram would seek (either internally or
externally) experienced expertise to
lead a short workshop (4-5 days max-
imum) which would focus on the
basic concepts and philosophy of
FSR/E combined with some selected
skills in diagnosis and design. The
goal of such a workshop is to develop
verbal ability on the subject among
a broad range of key individuals in
research and extension, as well as
management and administration. In
many respects, this type of workshop
can be designed for participants to
begin to develop a consensus on
what the national definition and
framework for FSR/E will be. It is
also a time for conflict to be con-
structively handled in order for
working and communication rela-
tionships to be established, especial-
ly across disciplines or even minis-
tries, where little crosss-over may
have occurred in the past.
2. Structured Networking
Following national level orienta-
tion (one or several workshops or
discussion sessions), a national pro-
gram can selectively and strategically

encourage future practitioners to visit
other more experienced practitioners
by sending them to participate in re-
gional, cross-country or international
FSR/E networking workshops and
symposia. Exchanges of personnel
with projects of greater depth of
experience can also be arranged. The
purpose of such structured exchange
is to solidify the context and applica-
tion of FSR/E by allowing practition-
ers to "see for themselves" or hear
others describe their experiences and
results. This helps people to under-
stand how FSR/E might serve their
own needs and purposes, and begins
to "internalize" the concepts and
3. Training in Methodological Skills
Following stage two, national pro-
gram leaders should be able to identi-
fy key individuals who are interested
and capable of beginning FSR/E work
within specific regions of the country
or in collaboration with ongoing
agricultural development activities.
It is best to begin small, and grow as
the level of experience develops and
the numbers of trained practitioners
increases. The selected individuals
should then receive intensive train-
ing in the methodological skills neces-
sary to begin farming systems activi-
ties. They can be sent to a regional
course, such as the one described
above, or the national program may
decide to hold such a course in
country in order to train a larger
"critical mass" at one time. Initially,
outside training expertise will be
needed to design and run such a
course. It is important to train at
least a small group to begin with so
that they can train the necessary
interdisciplinary teams to conduct
field work, and to be able to have
colleagues with whom to discuss
practical and theoretical aspects of
their work. Training only one or
two individuals and expecting them
to provide all of the training and
stimulation for a national program

is insufficient. Part of devleoping a
national training strategy is devel-
oping an indigenous capacity and by
institutionalizing the training process.
4. In-Service Training
After practitioners have begun to
implement FSR/E activities, based
on the skills learned in the training
course, they will need further train-
ing on specific subjects. These are
best dealt with in the context of
their actual ongoing work. For ex-
ample, novice practitioners may wish
to improve their skills in conducting
diagnostic surveys. Outside expertise
can be brought in to work with the
practitioners as they engage in an
actual diagnostic survey in a target-
ed area. Similarly, expertise can be
brought in again to deal with design
issues, statistical analysis, socioeco-
nomic analysis or the conduct of
on-farm trials. In many ways, this
stage resembles the "call system"
established by CIMMYT in East and
Southern Africa, and reinforces the
hands-on aspects of FSR/E. FSSP
has assisted projects in this fashion on
a number of occasions, for example,
by providing expertise needed for con-
ducting diagnostic surveys (Liberia
1984) and designing on-farm trials
(Gambia 1985, 1986).
The four stages described above
may or may not occur in the sequence
listed, however, a national program
should determine what kind of
sequence would best suit its specific
needs. This underlines the need for
national programs to be able to call
upon outside expertise to assist them
in assessing their training needs and
designing an appropriate strategy to
address them. Such expertise should
have a certain "neutrality" regarding
priority commodities or agroecologi-
cal settings in order to honestly help
national programs assess their needs
and not just those of donor or inter-
national entities.








Farming Systems Research and Extension is a methodology directed toward improvement in technology develop-
ment for limited resource farmers who, out of necessity, must distribute labor and investments across a variety of
production activities. This executive summary outlines the FSSP training strategy promoting understanding and
implementation of FSR/E.

Provide the basic foundations for development of indigenous training capacity in Farming Systems Research and
Extension (FSR/E) within National Institutions.

Initiate and implement regional and national practitioner level short-term FSR/E training in English and French
actuating a base for a long term training program for the region.

Develop Materials

* Slide Tape Modules
* Selected Readings
* Supporting Documents
networking papers
annotated bibliographies
* Training Units
On-Farm Experimentation
Livestock (pending
Economic Analysis
Management and Admin.
* Case Studies
Diagnosis Dominican
On-Farm Trials Paraguay
FSSP/Population Council
Burkina Faso
Eastern Caribbean

* Provide general orientation
to FSR/E concepts

* Focus on skillbuilding for
using methodological tools

* Focus on FSR/E practitioners

* Encourage flexibility for
tailoring training to
necessary to design

* Provide for participatory
or "hands on" activities

* Use real world circumstances

* Provide opportunity for
user input to the development
and modification of training

* Create mode for synthesis
and analysis of current
FSR/E methodologies

Deliver Shortcourses

* Regional Shortcourse,
English (3 weeks)
French (3 weeks)
Spanish (3 weeks)

* Individual shortcourses
on various aspects of
FSR/E offered on a buy-in
Possible topics include:
orientation to FSR/E
on-farm experimentation
analysis of on-farm ex.

* Encourage regional
participation from various
network organizations

* Contribute to national
training programs

* Build on previous
training activities

* Focus on building
methodological skills
necessary to conduct
FSR/E within a national
program context

Support Training Programs

* Train Trainers
prepared to offer
workshop on a buy-in

provide participant-
add-on opportunities
for non-trainer
resource persons for
in-service training
on an institutionally
supported basis.

* Advise National Programs
on a training strategy
on a buy-in basis.
provide orientation,
in-service and
methodological skills

provide networking

* Provide an institutional
continuing education basis
for an FSR/E continuing
education program:

consider needs for
training in methodology
and pedogological skills

emphasize planning and
organizational skills
necessary to design
and run FSR/E workshops

provide access to FSR/E
training resources
including materials
and trainers

* Encourage participation
in country and regional
level FSR/E training

Support a four stage national
training strategy
1. orientation
2. structure networks
3. methodological skill
4. in-service training




Derek Byerlee*

In most farming systems research programs, the
major emphasis is on adaptive research aimed at
providing improved technical information to farmers
for using available technology. This information is
usually summarized in the form of recommendations
for specified groups of farmers (often referred to as
recommendation domains (Tripp, 1986). Typically,
adaptive research teams meet annually to formulate
recommendations that represent the current state of
the art with respect to the "best" methods for pro-
ducing a particular crop or livestock enterprise.
While a good deal of attention has been given to the
appropriate methodology for conducting adaptive
on-farm research, very few guidelines have been devel-
oped for efficiently synthesizing and summarizing the
information generated by adaptive research for use by
extension and farmers. There has been a tendency to
condense the results of perhaps hundreds of experi-
ments conducted over several years into a single
"recipe" for cop production. This "recipe" approach
does little justice either to the resources invested in the
research or the information requirements of farmers in
an increasingly dynamic and complex agriculture. A
good deal of valuable experimental information is
wasted unless ways are found to more efficiently
organize its transfer to farmers.
It is of course unrealistic even in developed agri-
culture to make all experimental data available to
extension and farmers. The challenge for researchers
is to 1) synthesize their results so that the most im-
portant findings are available and 2) simplify these
findings in a form that can be readily understood by
extension and farmers. In this note we consider two
main outputs of this process; 1) prescriptive informa-
tion or recommendations and 2) auxiliary information
to help extension and farmers adapt the prescriptive
information to their own circumstances.
Defining a Recommendation
Despite the widespread use of the term "farmer
recommendation" it is rarely explicitly defined and in
fact, different implicit meanings are often attached to

*South Asian Regional Economist, CIMMYT, Islamabad. These issues
are discussed in more depth in a paper (Byerlee, 1986) available from
the CIMMYT Office, Box 1237, Islamabad, Pakistan.

the term. To most it means the "best" production
technique, often defined from a researchers' point of
view in terms of maximum yields or profits.
Here we adopt a working definition of a good
recommendation as that practice that farmers, given
their resources and objectives, would employ if they
had all the information available to the researchers.
Note that farmers' decision criteria rather than re-
searchers' criteria are the basis for selecting recom-
mendations. A recommendation assumes the current
resources and objectives of farmers and aims to allevi-
ate only one factor limiting productivity-the imperfect
information situation of farmers.
From this definition a number of other implications
1. A recommendation takes account of the existing
capital situation of farmers. If official bank credit is
in short supply (which it usually is), recommenda-
tions must often be formulated using a relatively
high minimum rate of return on capital (e.g. 100%).
2. A recommendation should only be made for in-
puts currently available to farmers. Availability is,
of course, a relative concept and researchers may
also want to provide information to policy makers
and input suppliers to improve the availability of
a given input.
3. A recommendation should provide new informa-
tion that promotes a change in current practices.
There is little point in recommending a practice
that is already widely used by target farmers.
4. A recommendation, since it refers to a future
event, must be defined in terms of a time horizon.
This means that a recommendation should em-
phasize a few priority changes that can be adopt-
ed by a farmer in the immediate future over two
or three crop cycles. Farmers do not adopt pack-
ages but because of capital constraints and a
cautious learning-by-doing approach, adopt im-
provements in a stepwise manner in a logical
sequence (Byerlee and Hesse de Polanco, 1986).
Intra-farm Variability and Conditional
In the early stages of an adaptive research program
the challenge is 1) to define relatively homogeneous
subgroups of farmers or recommendation domains
and 2) issue recommendations that are generally
applicable to these subgroups. The stratification of

farmers is a continuous process as research informa-
tion accumulates and recommendations are made
(Tripp, 1986). Stratification recognizes the substantial
variability characteristic of most target farm popula-
tions. This variability may be attributed to 1) variation
in crop response due to agro-climatic site and season
effects, 2) variation in prices paid and received by
farmers and 3) variation in the resource endowment
and objectives of farmers. There is always potential
to stratify farmers into smaller more homogeneous
groups of farmers in order to reduce between-farmer
variability due to all three sources. This process is
initially limited by resources available to research and
extension. However, as adaptive research programs
mature and research information accumulates a point
is reached where between-farm variability is less im-
portant than variability between fields and across years
within farms.
Farmers usually face substantial variation within
farms and in fact, traditional management practices
reflect adaptation to this variation. They often have
different soil or land types or location of a field in
relation to the village strongly influences management
and base fertility levels. Decision making for this
field-to-field variation within farms is sometimes
referred to as "strategic" since farmers know or can
measure this variation and accommodate it in their
management strategy.
In addition, farmers commonly face substantial
year-to-year variability, especially due to climatic and
pest hazards. They often take decisions in response
to this environmental variation, e.g. elimination of a
second nitrogen application with early season drought
or a decision to apply a pesticide based on observed
pest incidence in a field. This type of decision making
is sometimes referred to as "tactical". Both of these
types of decisions-strategic and tactical-provide
opportunities for conditional recommendations. That
is, a recommendation is made conditional on a given
event. For example, for wheat after maize apply 50
units of nitrogen and for wheat after sugarcane 100
units. Or if pest infestation exceeds a certain level
then spray.
Different types of conditional recommendations are
shown in Table 1. Both the conditioning variable and
the final recommendation may be discrete or continu-
ous. Examples of discrete conditional variables are
crop rotation, land type, soil type and pest presence or
absence. These are usually known or are easily observ-
able by farmers. Continuous conditional variables in-
clude soil test results, pest counts and recorded rainfall.
These require some effort and costs on the part of
farmers to measure. A common example is the thresh-
hold concept in pest management where pest scouting
is used to estimate pest population density and to pre-
dict pest damage. When predicted damage exceeds a
threshold determined by the cost of spraying a pesti-
cide, a recommendation is made to spray (e.g. Mumford
and Norton, 1984). Continuous conditional variables
may lead to continuous recommendations. The most

common example is soil testing to make fertilizer
recommendations, where recommendations are often
related to soil test results along a calibration curve.

Table 1: Examples of conditional recommendations

1. Fertilizer level x
crop rotation, soil/
land type (S)

2. Pest control x
pest presence/
absence (T)

1. Pest control x
pest count (T)

2. Irrigation x
soil moisture % (T)


1. Fertilizer
level x
soil test
level (S)

S = strategic decision. T = tactical decision. NA = not applicable.
The Appropriate Level of Recommendation Specificity
On the basis of the above discussion we can define
three levels of recommendations of increasing specificity.
Type I-General recommendations. A single recom-
mendation is made for all farmers in a
recommendation domain.
Type I I-Recommendations conditional on discrete
Type I I I-Recommendations conditional on contin-
uous variables.
More specific recommendations entail higher research
costs (more experiments), extension costs (more infor-
mation to transfer) and farmer costs (to measure con-
tinuous conditional variables). Hence, an adaptive
research program should invest in increasing specificity
of recommendations until the value of the additional
information exceeds the cost of generating, transfer-
ring and using the information. In developed countries,
farmers are increasingly provided Type III recommen-
dations (e.g. soil tests and integrated pest management
based on pest scouting) and most studies suggest that
the value of the additional information exceeds the
cost (Byerlee, 1986). In developing countries, major
emphasis must still be placed on useful Type I recom-
mendations. Considerable scope exists for Type II
recommendations conditional on variables such as land
type and crop rotation, although costs of transferring
this information may be higher for farmers lacking
literacy and numeracy skills. Many reserach systems
have jumped to Type II recommendations such as
fertilizer doses based on soil tests even though the
appropriate supporting services such as soil test labora-
tories are not available or are not reliable.1
(continued on next page)

1 Although soil testing laboratories are available in many countries,
most researchers complain of long delays and inconsistent results.
Hence, it will be a long time before small farmers have ready access
to soil test results.



The definition of recommendation domains and
conditional recommendations must be made in terms
of variables that are easily identifiable by extension
workers. In practice, a good deal of variability between
farmers is due to factors, especially socio-economic
circumstances such as access to capital, that are im-
portant in farmers' decisions on technology, but which
can not easily be employed in the field by extension
agents as delineating variables. This type of variability
is often greater in developing agriculture.
In addition, in a dynamic agriculture with a con-
stantly evolving technical and economic environment,
it is expensive to continuously update and transfer very
specific recommendations. Hence, recommendations
are likely to rapidly become outdated and farmers
need to be able to adapt technology to changes in
the environment.
For these reasons recommendations, whether
general or specific, are only guidelines which farmers
must adapt to their own individual needs. Adaptive
research programs usually generate a good deal of
information that is not incorporated into the recom-
mendations but which would be valuable to farmers
in their own adaptation of recommendations. Ex-
amples of this type of information are:
1. Economic data on costs and returns. For example,
the net benefit curve (see Perrin et. al, 1976)
provides a set of efficient treatments for each
level of expenditure and at different minimum
marginal rates of return.2 Optimum combina-
tions of inputs (e.g. nitrogen and phosphorous)
can be easily derived for a given expenditure
2. Data on the sensitivity of the response of recom-
mended technology, especially to the timing and
method of application. In the course of conduct-
ing adaptive experiments, researchers learn a great
deal about factors that affect the efficiency in
using a given input under farmers' conditions.
Information on downside risks (e.g. phytotox-
icity from late herbicide applications) could be
particular important in speeding adoption.
3. Information on feasible sequences for the adop-
tion of two or more technological components.
As noted, farmers do not usually adopt a package,
but proceed a step at a time. However, there are
often interactions between technological com-
ponents which indicate that one component
should be adopted before another (e.g. weed con-
trol before fertilizer) while the reverse sequence
might not be economic.
Beyond this auxiliary information from adapative
research, farmers can more effectively adapt recom-

2The net benefit curve is the set of undominated treatments and is equi-
valent to the expansion path in conventional production economics.
'The feedback of information from on-farm research to set priorities for
applied on-station research is also an important role.

mendations if they understand basic principles of
crop production and have basic technical and man-
agerial skills. For example, knowledge of nutrient
composition, potential carryover effects and symptoms
of nutrient deficiencies should help farmers adapt
fertilizer recommendations to their own needs. Indeed
some recommendations may depend on farmers' under-
standing of principles if they are to be widely adopted.
Examples include changes in variety to avoid break-
down of pest resistance as new races of a pest evolve,
or use of complex pest management practices that
depend on survival of beneficial insects. In addition,
the technical efficiency with which a recommendation
is used will depend on farmers' skills in calculating
doses, identifying pests, calibrating sprayers etc.
In many cases, extension services themselves are
severely deficient in their basic understanding of
production principles and technical skills. Initially
researchers may want to concentrate on training of
extension agents and providing them a wider range of
auxiliary information, so that they can help farmers
more effectively adapt and use the new technology.
Over time, it is desirable that farmers themsevles
acquire this knowledge and related skills. In a changing
environment this knowledge and related skills will also
depreciate more slowly than prescriptive information.
This implies that extension give more emphasis to an
educational role as opposed to its traditional role in
communicating messages.
Farming systems research is largely a process of
generating better information for farmers to improve
their productivity.3 There has, however, been a tend-
dency to summarize this information into "recipes"
for crop production that does little justice to the com-
plexity of most farmer decisions or to the large amount
of useful information generated in on-farm trials.
Farmers can be provided access to this information
through more specific conditional recommendations.
However, in the longer run the challenge is to provide
farmers the wider range of information, knowledge of
production principles and technical and managerial
skills to evaluate and adapt new agricultural tech-
nologies to their own needs.E

Byerlee, D. "Essential Conceptual Issues Related to Adaptive
Research, Farmer Stratification and Extension Advice",
Draft Paper, CIMMYT, Islamabad, Pakistan, 1986.
Byerlee, D. and E. Hesse de Polanco. "Farmers' Stepwise
Adoption of a Technological Package: Evidence from the
Mexican Altiplano". Amer. J. Agric. Econ., Aug., 1986.
Mumford, J. D. and G. A. Norton, "Economics of Decision
Making in Pest Management". Ann. Rev. Entomol. 29
(1984): 157-74.
Perrin, R. K. et al., From Agronomic Data to Farmer Recom-
mendations: A Training Manual. Inform. Bul. No. 27, El
Batan, Mexico; CIMMYT, 1976.
Tripp, R., "Some Common Sense About Recommendation
Domains", Farming Systems Support Project Newsletter,
4 (1986); 1-3.


New challenges are facing people
who are concerned with agricultural
development in less developed coun-
tries. Despite significant increases in
production of some crops, such as
rice and wheat, more people than
ever today are suffering from malnu-
trition. The gap between experimen-
tal and farmers' yields has prompted
scientists and development workers
to seek ways to develop agricultural
technology for farm households with
limited resources. Although new de-
signs for rapid rural appraisals and
on-farm experimentation have im-
proved the perception of farmers'
constraints, it is apparent that tech-
nology appropriate to agricultural
intensification in diverse tropical
environments is not readily available.
As a result, there is a growing aware-
ness of the need for a holistic and in-
terdisciplinary perspective to develop
ecologically sustainable and socially
desirable farming and cropping sys-
tems. Also, change in technology
provides only part of the answer and
must be complemented by suitable
development policies in such areas as
pricing, marketing and infrastructure.
The series Development Oriented

Research in Agriculture addresses
these issues. It intends to provide an
opportunity for exchange of experi-
ences and ideas among policy makers,
agricultural scientists and develop-
ment workers in less developed and
western countries.
An average of one or two volumes
will appear annually. The volume,
Cassava in shifting cultivation, a
systems approach to agricultural
technology development in Africa
by L. O. Fresco was released in
November 1986 (see insert). Vol-
umes in preparation include: the
relationship between extension and
agricultural research; linkage between
commodity and farming systems
research; a case study on sorghum
agronomy in East Africa; and crop-
livestock interactions. Readers are
encouraged to submit suggestions for
future publications. Address corres-
pondence and order information re-
quests to:

Royal Tropical Institute
Publications Department
Mauritskade 63
1092 AD Amsterdam
The Netherlands


The following are new books re-
lated to FSR/E that have come to
our attention:
Benor, D. and Michael Baxter.
1984. Training and Visit Extension.
World Bank Publications, 214 pp.
Cernea, M. M., John K. Coulter,
and John F. A. Russell. 1985. Re-
search-Extension-Farmer: A Two-
Way Continuum for Agricultural
Development. World Bank Publica-
tions. 192 pp. $14.00
Elz, Dieter. 1985. Planning and
Management of Agricultural Re-
search. World Bank Publications.
160 pp. $12.50
Francis, C. A. 1986. Multiple
Cropping Systems. MacMillan Pub-

lishing Company: New York, N.Y.
400 pp. $37.50
Hansen, A. and D. E. McMillan.
1986. Food in Sub-Saharan Africa.
Lynne Rienner Publishers, Inc.:
Boulder, Colorado. 410 pp. $14.95
Hildebrand, P. E., ed. 1986. Per-
spectives on Farming Systems Re-
search and Extension. Lynne Rienner
Publishers, Inc.: Boulder, Colorado.
167 pp. $16.00
ICRISAT. 1986. Farming Systems
Principles for Improved Food Pro-
duction in the Arid, Semi-arid, and
Humid Tropics. 36 pp. $5.10
Jones, J. R. and Ben J. Wallace,
eds. 1986. Social Sciences and Farm-
ing Systems Research. Methodolog-
ical Perspectives on Agricultural De-

velopment. Westview: London and
Boulder, Colorado. 285 pp. $21.00
Moock Lewinger, J. 1986. Under-
standing Africa's Rural Households
and Farming Systems. Westview Press,
Boulder, Colorado and London. 234
pp. $18.85
Remenyi, J. V., ed. 1985. Agricul-
tural Systems Research for Develop-
ing Countries. Australian Centre for
International Agriculture Research,
ACIAR Proceedings No. 11. 189 pp.
Canberra, Australia.
Richards, Paul. 1985. Indigenous
Agricultural Revolution: Ecology
and Food Production in West Africa.
Westview Press: Boulder, Colorado
and Hutchinson: London and Mel-
bourne. 192 pp. $28.00


Although aggregated data for
Africa show a declining food avail-
ability per head, these do not ade-
quately reflect the diversity in
performance between crops and
regions. This volume presents a case
study from the Kwango-Kwilu region
in central Zaire. In this area, cassava
production has increased consider-
ably in the last thirty years and has
kept pace with or even surpassed
population growth, despite socio-
economic and agronomic disincen-
The relevance of this study lies in
its detailed analysis of changes in
shifting cultivation as well as in its
method of analysis. It draws upon
ecological system analysis and, to
a lesser extent, on farming systems
research, and presents a systems
framework that allows the integra-
tion of technical and socio-eco-
nomic aspects of crop production
which has wide application.

On-Farm Research in

Morogoro District

Manasse T. Mtoi* and Akwilin J.P. Tarimo**
The Farming Systems Research (FSR) approach at
the Sokoine University of Agriculture (SUA) has
recognized the need to improve the existing farming
system in two distinctive areas of Morogoro District:
the Uluguru Mountains and the Northern Plains.
These sites were selected after identifying farmer
circumstances and their production processes. These
include cropping systems, resource availability and
utilization pattern, farmers' priority objectives and
farm-household interaction. The diagnostic survey to
identify these components was done in 1979 through
1981 in the two sites. On-farm experimentation began
in 1981/82 and 1982/83 seasons in the Uluguru
Mountains and the Northern Plains respectively.
The project is jointly run by the International De-
velopment Research Centre (IDRC) of Canada and
the Sokoine University of Agriculture. This project
started in 1981 and is expected to wind up in 1989.
The second phase of the project is expected to take off
in November 1986.
Research scientists are becoming more conscious of
the fact that the impact of a new technology is meas-
ured not by its excellence in experimental plots but
by the extent to which it is adopted by farmers on
their farms.
Emperical evidence shows that when agricultural
research is undertaken on crops without high levels of
environmental control, substantial yield variations can
be expected between the experimental stations and the
farmers' fields. Further, there also exists a gap between
the availability of improved technology from research
stations and its actual application on the farms. Possi-
ble reasons for these gaps in the yield between the
experimental station and the farms, and between the
availability of the technology and its adoptions, are
vast and many. Although there is no agreement about
the nature and relative importance of such causes,
there is general consensus that some technologies
being evolved by the agricultural scientists do not
coincide with farmers' objectives and may not fit into
their social, economic, and production constraints
especially in view of their attitudes, risk potential,
and limited availability of most production inputs
like labor, capital etc. (2)
*Assistant Lecturer and Farming Systems Research Economist,
Department of Rural Economy, Sokoine University of Agriculture,
P.O. Box 3007, Morogoro, Tanzania.
*Assistant Lecturer and Farming Systems Research Agronomist,
Department of Crop Sceince and Production, Sokoine University
of Agriculture, P.O. Box 3005, Morogoro, Tanzania.

The objective of the project include:
*To study the present farming systems in Morogoro
District in order to identify the constraints to
increased production;
*To design technologies to be tested in adaptive
research experiments with farmers;
*To evaluate suitability and acceptability of these
recommended technologies;
*To train B.Sc. and post-graduate Tanzanian stu-
dents in the concept and methodology of FSR.
The concept and philosophy of the farming systems
approach to research and agricultural development is
still evolving and therefore, a conventional wisdom is
slowly emerging for solving methodological and im-
plementation problems. The most important consider-
ation is that small farmers with capital scarcity, risk
avoidance objectives and cautious learning process,
rarely make drastic changes in their farming system.
Inevitably a methodology should base a research
strategy conforming to farmer's characteristics and at
the same time ensuring effective linkages between
physical, biological and human factors.
The methodology adopted at SUA is consistent with
the conventional FSR attributes. Because of the inter-
disciplinary nature of the approach and the availability
of experts from different disciplines at the same loca-
tion, SUA is considered as a logical and ideal place to
test this methodology. The methodology outlined in
Figure 1 is dynamic, iterative and interdisciplinary and
has attempted to characterize FSR in the following
broad criteria.
The Descriptive or Diagnostic Stage. The actual
environment to identify constraints farmers face and
to ascertain the potential flexibility in the FSR in terms
of timing of farm operations, slack resources, etc. This
stage was undertaken in 1979 through 1981 in the
Uluguru Mountains and the Northern Plains by social
The Design Stage. A range of strategies pertinent
and relevant in dealing with the constraints delineated
in the descriptive or diagnostic stage were screened.
Inventory of available component technology derived
from research activities conducted at SUA and other
research institutes has been instrumental in streamlin-
ing the design stage. Farmers' knowledge and experience
were also incorporated.
From the identified constraints the following
hypotheses were formulated:
1. Low plant population and planting configuration
could contribute to low yields of cassava, pine-
apples and sorghum.
2. Poor weed control practices could contribute
significantly to seasonal low yields of the crops
identified in number 1, above.
3. Late planting causes yield variation.
4. Comparative benefits and costs analysis between
the traditional and alternative farm production

systems would differ given different conditions in
terms of management and efficiency of resource
allocation. Under the presumption that resources
are being utilized more efficiently on alternative
systems, the system would provide a form of
maximizing net farm income per unit area.
5. The impact of alternative farm production sys-
tem would reduce income variability and hence
farmers' risks toward farming.
The Testing Stage. A few promising strategies arising
from the design stage were examined and evaluated
under farm conditions to ascertain their suitability for
producing desirable and acceptable changes in the
farming system. This stage involved initial trials at the
farm level with joint researcher and farmer participa-
tion, then, farmers' testing with total control by
farmers themselves.
The Extension Stage. The strategies identified and
screened during the design and testing stage would be
implemented. This is expected to be done in the second
phase starting in the 1986/87 cropping season.

Figure 1. Farming Systems Research Methodology, Morogoro District. a

Contemporary FSR has applied objectives to develop
and disseminate appropriate technologies that increase
agricultural productivity and improve the standard of
living of small holders. In order to achieve these objec-
tives the FSR team differentiated groups of small
holders, identified the specific needs of these different
groups and developed potential technologies which
represented real improvements over existing practices.
The recommendation domain was identified in two
locations, Kinole ward in the Uluguru Mountains and
Melela ward in the Northern Plains. Both sites exhibit
totally different farming systems. In each case the FSR
team divided the area into sub-areas with relatively
uniform characteristics.
Research sites, representative of the selected sub-
areas were then selected. For example, in Kinole ward,
there were three sub-areas comprised of Tonya Juu,
Kisambwa and Tonya.
Selection of sites was made by considering the
potentiality of the areas with respect to FSR. Other
criteria used in the selection of sites
included accessibility to the area. A
small area to manage, in terms of
ing Farm System transport and other facilities, was
favored over an ambitiously large area.
nt Technology The target group farmers were
er Circumstances chosen on the basis of common pro-
er Objectives duction patterns and farming practices.
er Priorities
er Priorities Such a classification was considered
action and adequate for identifying problems and
umption Patterns opportunities of sufficient magnitude
able Resources to justify the research effort. Village
iEconomic leaders participated in the selection of
utions the farmers. A total of 16 farmers in
the Uluguru Mountinas and 15 farmers
in the Northern Plains agreed to pro-
vide land and participate in the farm
Figure 2 (see following page) shows
the mechanisms with which the
Farming Systems Research Project at
SUA operates. At one end there is the
University Administration which is in
charge of the overall project and
particularly responsible for adminis-
tration of funds.
Principal investigators include the
project leader and the assistant project
leader who are responsible for research
activities and financial expenditure.
The interdisciplinary team decides on
relevant research activities and is
responsible for the design of experi-
ments and in the discussion of research

a Adapted after Mtoi, M.T.T. and Anandajayasekaram, P.

The research assistants provide day-to-day feedback
to researchers, extract information and monitor the
experimental plots. The village assistants manage plots
and assist in data collection.
The project managed to train two principal investi-
gators and one statistician in diagnostic phase and ex-
perimental phase workshops held in Harare, Zimbabwe
in September, 1984 and February, 1985 respectively.
The two principal investigators also attended an in-
country training program held in Moshi, Tanzania in
May, 1986.
The nature and scope of problems identified depend
on the methodology adopted. Methodology for under-
taking adaptive research as we have seen, is still going
through a period of evolution. Not surprisingly perhaps,
there are often considerable differences in opinions on
firstly, prioritizing identified problems and secondly,
how they should be dealt with.
Some problems obviously would require institutional
intervention. The major problem areas identified at
SUA's FSR activities have been
classified in terms of biologcial, Agriculture C
socio-economic and physical Engineering Sc
problems. Biological problems
include poor soil fertility conser-
vation methods, low plant population and poor plant-
ing configuration, pests and diseases, timing of farm
operations and use of inferior quality varieties. Socio-
economic problems included land scarcity (in the
Uluguru Mountains), low cash flow, labor scarcity,
poor infrastructure, and off-farm activities, partic-
ularly in the Northern Plains. Insufficient moisture
in the Northern Plains (semi-arid area) was found to
be one of the major physical problems.
The on-farm research project at SUA has widened
the knowledge and experiences of farmers and
researchers at the Sokoine University of Agriculture.
Experiments conducted in the areas as possible solu-
tions to the identified constraints have produced very
useful results.
For example, weeding experiments in farmers'
sorghum fields showed that about 40 percent of the
farmers crop was lost to weeds, particularly the Striga.
Farmers were able to weed less than 50 percent of the
total cultivated area; such results were considered
useful in designing experiments in the second phase.
Experiments on "some recommended practices" in
crop production proved useful, such as intercropping
of cereals with legumes, proper spacing for cereals and
legumes, as well as for cassava and pineapples. Com-
parative studies on monoculture and intercropping,
fertilizer and plant protection chemicals against pests
and diseases also proved both useful and practical to
In the Uluguru Mountains a cassava technology with
regard to triangular spacing 1.0m x 1.0m x 0.75m was

found to be superior than farmers' traditional tech-
nology. The average yield in the alternative technology
was found to be about 32 percent more, whereas
associated returns increased by about 33 percent.
The coefficient of variation, defined as the standard
deviation of net farm income divided by average of net

Figure 2. Organization Chart for On-farm Research at SUA.

University Administration

t F
Principal Investigators

Adaptive Research Team

S Other
Animal Related
Science Depts.

Research Asistanfl

income was used to measure income variability in the
two systems of production. The coefficients were 46
and 29 for the traditional and alternative systems
respectively. The finding of relatively high income
variability in the traditional system suggests that
growing of cassava and pineapples using traditional
methods reflects a greater degree of vulnerability to
the farming system.
Importantly the on-farm research at Morogoro
exposed researchers to the complexity of combined
research efforts, an approach which was lacking prior
to the introduction of the FSR project. Through the
combined effort, researchers were able to identify
farmers' priorities and problems. The concept of shar-
ing and comparing data was experienced in the inter-
disciplinary approach. The farming system research
project is a common pool for various data pertaining to
individual specialities which were intended for devel-
oping the small farmers in Tanzania.
Coordination of research and joint publication of

research findings were among other significant achieve-
ments experienced by researchers of farming systems
at SUA.
Another important experience gained by the re-
searchers at SUA was the art of doing research with
farmers. Farmers were not only co-researchers but
were also among the research variables to be analyzed.
The interactive nature of the FSR methodolgy pro-
vided for free feedback mechanisms for researchers to
evaluate and re-design on-farm experiments. From this
systems approach, research and dissemination of tech-
nology occurred at the same time. It has thus been
tentatively agreed at Morogoro by the farming systems
researchers that the FSR approach should prove a real
pathway for discovering small farmers' developmental
interests. Such a discovery would enhance the rate of
development of compatible agricultural production
The Farming Systems Research Project at SUA is
now entering into its second phase. This phase is
important since some of the tested technologies in
the first phase could not produce conclusive results.
Secondly, the developed technologies in phase one
need further evaluation in order to assess their compat-
ibility with the small farmers. Thirdly, time is needed
to disseminate the developed technologies to other
farmers in the two recommendation domains. Fourthly,
the second phase will be used to design further tech-
nologies for continued on-farm research.
Phase two of the FSR project will concentrate its
activities mainly in the Northern Plains because the
Uluguru Mountains area has some physical infrastruc-
tural problems impairing accessibility during the rainy
In brief, on-farm research activities during phase II
will focus mainly on cropping systems. The experi-
ments will be directed toward:
1. developing efficient methods of controlling the
witchweed (Striga hermonthica) on sorghum;
2. multiplicaton of acceptable early maturing
varieties of sorghum, maize, pigeon peas, ground-
nuts and greengram for distribution to farmers;
3. soil fertility studies, and
4. time of planting studies.
SUA's on-farm research has identified two key
scenarios which require institutional support, Firstly,
the relationship between on-farm research and exten-
sion has to be well recognized, and secondly, the
institutionalization of on-farm research has to be done.
The first scenario is extremely important in that SUA's
research is entering a second phase where the role of
extension agents is of paramount importance.
Perhaps limited budgets and manpower are obvious
impediments in the process of institutionalization of

FSR activities. The conceptual and philosophical appeal
of FSR has allowed international agencies such as I DRC
and USAID to speed up this process in Tanzania. This
scenario poses two important questions: what will
happen to FSR programs when the donor agencies
leave? Are the national institutions going to maintain
the same level of efforts in terms of resources to con-
duct on-farm research?
Institutionalization of FSR will obviously oblige
the Government through the appropriate Ministries
to allocate the necessary resources if on-farm research
activities should continue. Budget cuts by donor
agencies frequently scale down anticipated on-farm
research activities in the collaborating institutions.
The shift in recent years to incorporate on-farm
research into agricultural research is considered as a
timely and necessary step in addressing the funda-
mental questions of rural development. The small
farmer should still be an important client for new
technology developed for the purpose of increasing
basic food as well as cash crop production.
Very promising results have been achieved in the
on-farm program at SUA. Specific packages of im-
proved technology have been identified and are being
tested for compatibility and acceptability by farmers
in respect of their socioeconomic, cultural and pro-
duction constraints.
The future of SUA's on-farm research programs
must lie within national agricultural development
programs. Strong Government support will be im-
perative in implementing FSR production programs
in the near future. E

Anandajayaseram, P., Farming Systems Approach to Research
and Development in Tanzania". Paper presented at the meet-
ing of Tanzania Agricultural Economic Society, Dar es
Salaam, 1980.
Hildebrand, P. E., "Design and Analysis of on-farm Agronomic
Trials". TMS 501, FSSP June, 1983.
Oram, P. A., "The nature and Extent of the Problem: An over-
view of Research and Extension Needs". Paper presented at
the Workshop on Financing the Recurrent Costs of Agricul-
tural Services. The Hague, ISNAP, May 20, 1983.
Mtoi, M. T. T., "Comparative Analysis of Resource Allocation
under Risk and Uncertainty in Peasant Farming in Morogoro,
Tanzania". M.Sc. Thesis Submitted to the University of
Guelp, Canada, April, 1984.
Mtoi, M. T. T., and Tarimo, A. J. P., "Bringing National Policy
Objectives to Bear in the course of the locally Oriented
Farming Systems Research Process". Paper presented at the
Regional Farming Systems Research Workshop held at
Egerton College, Kenya, August 19-23, 1985.
Tarimo, A. J. P. and Mtoi, M. T. T., "An overview of Farming
Systems Research Activity at Sokoine University of Agri-
culture". Paper presented at a Farming Systems Research
Workshop held at AICC, Arusha, May 20-25, 1985.
Lugole, J. S., Mtoi, M. T. T. and Tarimo, A. J. P. "Adaptive
Research in the Uluguru Mountains and the Northern Plains
of Morogoro District". Paper presented at a National Farm-
ing Systems Research Workshop, AICC Arusha, November
12-13, 1985.




"In the late 1960s, the Govern-
ment of Guatemala conducted a
comprehensive assessment of its rural
areas. The assessment indicated that
food production was just barely
keeping pace with growing demand
and that rural incomes and farmer
productivity were stagnating. Mini-
mal increases in production had been
achieved primarily by increasing the
land area devoted to agriculture.
Particularly affected were the bean
and maize staple food crops. Although
bean production had doubled between
1960 and 1970, total acreage had
increased almost three times. During
the same period of time, yields of
maize, the most important food crop
in Guatemala, had scarcely increased
at all. While the export subsector of
agriculture contributed $211 million
of foreign exchange in 1972, well
above the $21 million agricultural
import bill, the country still had to
import maize and beans. Increasing
amounts of foreign exchange were
being allocated to purchase basic food
imports. Complicating the problem,
it was observed that the availability
of sufficient arable land was becom-
ing a major constraint to maintaining
needed production levels". (McDer-
mott and Bathrick, 1981, p. 1)
Donors are anxious to evaluate the

*Professor, Food and Resource Economics
Department, University of Florida, 2126
McCarty, Gainesville, Florida 32611.
**Editor, FSSP and Center for Tropical Agrri-
culture, University of Florida, Gainesville,
Florida 32611.

Peter Hildebrand* and Steve Kearl**

impact of their investments, including
in recent years an increasing number
of farming systems projects designed
to help solve the kind of problems
illustrated above. An FSSP inventory
undertaken in 1984 identified nearly
300 farming systems research projects
and/or projects containing farming
systems components. Most had been
in operation for less than five years.
Also, most of these projects were
pilot projects or involved activities in
only a limited portion of the country
in which they were located. Because
of year-to-year variability it is diffi-
cult to make an evaluation of a tech-
nology development and diffusion
project in agriculture. It is addition-
ally difficult if the project is not
carried out on a national scale.
The impact of a national farming
systems program is clearly evident
in the case of the Guatemala Insti-
tute of Agricultural Science and
Technology (ICTA). This national
institute was created in 1973 by the
Guatemalan government with the
financial support of USAID, which
supplied transportation, expatriate
expertise, and operating funds, and
The Rockefeller Foundation, which
supplied additional expatriate exper-
tise. Primary support, however, came
from the Guatemalan government
which provided 90 percent of the
professional, administrative and tech-
nical personnel over the first six
years and an even higher proportion
after that. The government also pro-
vided the facilities, much of the trans-
portation and most of the operating
funds. The Institute has also been

supported in its efforts through
collaboration with CIMMYT, CIAT,

Challenge Facing ICTA
The original emphasis in the insti-
tution was to increase the production
of basic grains which were produced
predominately on small farms. Ap-
proximately two-thirds of the six
million people in the country lived
on small farms and they produced
two-thirds of the basic grain produc-
tion. The primary objective of the
government was to achieve self
sufficiency in the most important
basic food grains-maize, and beans
-and increase the production of the
others-sorghum, rice and wheat. A
later objective was to help small farm-
ers diversify production to reduce
their susceptibility to price fluctua-
tions and also to enhance the earnings
of foreign exchange from exports.
During the first decade of ICTA's
existence, the country not only
suffered the effects of spiraling
petroleum prices, rapid inflation and
currency devaluation, but also was
involved with fighting a guerrilla
insurgency. Nevertheless, the goals
upon which the institution was based
have been surpassed.
A 1985 ICTA publication states
(translated), "Even though only a
few years have elapsed since ICTA
was established, it has made some
impressive progress in maize, bean
and rice production; so much that
the country has achieved self suffic-
iency in these crops. Appropriate
technology has also been developed

for sesame, soybeans, sorghum, wheat,
potatoes, onions, tomatoes, broccoli,
cauliflower, English peas and snow
peas, and a stable export market has
been established for the production
of melons. It is also expected that
table and wine grape production will
soon be a reality because two varieties
are being evaluated in validation
trials. ICTA has also designed simple
and inexpensive storage structures
which permit the small producer of
potatoes and apples to store their
products for 90 days after harvest
and sell when prices are higher. These
structures are spreading rapidly
among farmers because they are
simple, low cost and effective".
(ICTA, p. 7)
Obvious Success
Figures 1, 2 and 3 show the prog-
ress that has been made in the yield
and production of maize, beans and
rice since the establishment of
ICTA. During the period of time
shown on the figures, the population
increased from approximately six
million to seven million, or nearly 17
percent. Maize yields (Figure 1) in-
creased from approximately 1120
kg/ha to 1570, an increase of 40
percent. Coupled with an increase
in area planted, production increased
from 670,000 tons to 1,080,000, or
61 percent. This represents a per
capital increase in maize production,
the most important basic food crop
of the country, of approximately 38
percent. This is a dramatic increase
for a crop that is raised predomin-
ately on small farms, even during a
time of economic and civil turmoil
in the country.
During the same period of time,
bean production, the main source
of protein, particularly for the poor
and for the small farmers, increased
from 58,000 to 90,000 tons based
mostly on an increased yield of
270 kg/ha, from 700 to 970 kg/ha
(Figure 2). This is a per capital in-
crease of 33 percent on a crop that is
also predomimately grown on small
holdings and very often in association
with maize and/or sorghum.
Although there is some upland
rice grown on small holdings, most
rice (Figure 3) is irrigated and grown
on large farms. Production has in-

creased 150 percent resulting in an
increase of more than 50 percent
in area but also nearly a 70 percent
increase in yield. The result more
than doubles per capital production
of this grain.
Role of Farming Systems Research
During the first half of the decade

Z 1600
z 1400

o 800

Z 600


74 75 76 77 78

reported in these data, ICTA was
pioneering Farming Systems metho-
dology. During this evolution, mis-
takes were made and corrected;
revisions were the results of team
discussions with everyone participat-
ing. National Coordinators met on a
weekly basis and regional reviews




79 80 81 82

..................... PRODUCTION, METRIC TONS
---. AREA

FIG. 2

Trends in the production area and per hectare productivity (yield) of maize in
Guatemala, 1974-1982.

74 75 76 77 78 79 80 81 82

.................... PRODUCTION, METRIC TONS

Trends in the production, area and per hectare productivity yield of common
beans in Guatemala, 1974-1982. (The drop in yield in 1976-77 was due to
golden mosaic virus.)

and evaluations were held every year
to program the following year's work.
The institute initiated programs in
three of the country's seven regions
and gradually increased its area of
influence by adding one region at a
time. The most impressive advances
have been made in the area of genetic

Trends in the production, area and per
Guatemala, 1974-1982.

w 50-
z 40
< 30
u 20-

S10o -

FIG. 3


improvement, but improved agro-
nomic practices have also been
The examples of ICTA in Guate-
mala demonstrates that the Farming
Systems approach is feasible on a
national scale and that it can be ef-
fective even under adverse conditions.

hectare productivity (yield) of rice in

*/ \/%.h.


- 2000


Production increases and self-suffic-
iency in basic food grains have be-
come a reality. For Guatemala, the
ICTA Farming Systems approach has
directly contributed to meeting
national agricultural development
I n other ways the I CTA experience
has also contributed greatly, both to
farming systems methodology and
the growing body of worldwide
experience. ICTA must be credited
with having contributed directly and
indirectly to the FSSP and its training
materials development. It has contri-
buted to the human resource capa-
bility which, over time, has found
former ICTA personnel in other new
and developing farming systems
projects and programs. This shared
experience is both a compliment to
ICTA and a contribution to the state
of the art.E

ICTA. 1985. Logros y contribuciones a la
agriculture national. Guatemala.
McDermott, J. K. and David Bathrick.
1981. Guatemala: Development of the
Institute of Agricultural Science and
Technology (ICTA) and its impact on
agricultural research and farm produc-
tivity. U.S. Agency for International De--
velopment, Washington, D.C.

Country Orientation Notebooks for Selected Countries in Africa

The Center for African Studies
and the International Programs of
the Institute of Food and Agricul-
tural Sciences at the University of
Florida have produced a series of
Country Orientation Notebooks for
selected countries in Africa: Burkina
Faso, Burundi, The Gambia, Liberia,
Mali, Rwanda, Sierra Leone and
Zaire. The notebooks focus on agri-
cultural development and are inter-
disciplinary in content, designed for
pre-departure orientation for con-
sultants assigned to agricultural
projects sponsored by international
development agencies and private
voluntary organizations. The note-
books are also intended to serve as
a resource for instructors and stu-
dents to use in studying agricultural

development in specific African
Each notebook contains 300-500
pages of material in a ring-binder
format including:
Introduction by a country
*General Information
(Post Report)
*Historical Survey
*Current News Articles
(African Journals)
*Statistical Information
(World Bank, FAO, AID)
*Agricultural Articles
*Agricultural Development
(country strategy reports,
national plans, congressional

(annotated section on agri-
*Slide-tape module depicting
country and its agriculture.
The cost of each notebook is $175
(U.S. prepaid). After the initial
purchase of a notebook, semi-annual
updates of current news clippings,
reports and articles is available for a
fee of $25 (U.S. prepaid) per update.
Both prepaid prices include shipping
and handling. For further information
or to order notebooks, contact:

Center for African Studies
Orientation Notebooks
470 Grinter Hall
University of Florida
Gainesville, Florida 32611

I I I I I I I I I 0
74 75 76 77 78 79 80 81 82
.................... PRODUCTION, METRIC TONS


Report of a Workshop on

Non-Traditional Experimental Designs

for On-Farm Systems Experiments

Clive Lightfoot', John S. Caldwell2
Wesley Kline1 and Neal Thomas3

During the sixth annual farming
Systems Research and Extension
Symposium held October 5-8, 1986,
at Manhattan, Kansas, a workshop
was held on non-traditional designs
for on-farm systems experiments.
The workshop began with a four hour
pre-symposium session October 5
and concluded with a second two
hour session October 6. This report
describes the purpose, organization,
results, and recommendations of the
Workshop purpose
This workshop built on a workshop
held at the 1985 symposium. Over
the past five years, Farming Systems
Research/Extension (FSR/E) practi-
tioners have made slow progress in
the development of experimental
designs for FSR/E. Each year, the
annual symposium has covered much
of the same ground. In the 1985
workshop, Barker and Lightfoot
(1985) outlined the problems of
traditional on-farm experiments and
presented new directions for success-
ful experiments based on a world-
wide survey of 41 FSR/E projects.

'International Agriculture Program, Cornell
University, Ithaca, N.Y. 24853, U.S.A.
2 Department of Horticulture, Virginia Poly-
technic Institute and State University, Blacks-
burg, Virginia 24061, U.S.A.
jBox 58 RRI. Mallory Town, Ont. KOE 1RO.

The purposes of this workshop
were three:
1. To identify experiences in the field
with three types of non-tradition-
al designs developed in response
to the new directions presented in
a. Independent farm experimenta-
These are cases where the re-
search team has discovered
independent farmer experimen-
tation and developed techniques
to analyze and use that experi-
mentation in formal technology
b. Flexibility in researcher-planned
These are designs that are flexi-
ble enough to accommodate
differences among farms and
farmer modifications not only
before establishment of the
experiment, but also during the
course of the experiment after
establishment, yet still produce
data sets that can be subjected
to predictive analysis.
c. Analysis of systems interactions.
These are designs that use tech-
niques to monitor both biolo-
gical and social effects of the
trial on all or many of the
system components and evalu-
ate the trial results in terms of
trade-offs among systems inter-
actions using both the biological
and social data.

2. To develop a network of practi-
tioners for continued exchange of
information and ideas on new
approaches to on-farm systems
3. To identify needs for improve-
ments in the on-farm experimenta-
tion training materials and work-
shops developed by the FSSP, and
examples from the field to use in
the training materials and work-
This report focuses on the first and
third purpose. A list of practitioners
and ideas for a network will be pre-
sented in a later newsletter issue.
Session organization
In the first session (October 5), a
two-page survey of experience in four
topic areas was administered to 45
participants, to identify potential
"key informants" and divide the
participants into four groups for dis-
cussion of each topic area. The four
topic areas were based on the new
directions suggested by the findings
of the previous survey (Barker and
Lightfoot, 1985). The first three topic
areas corresponded to the three types
of experiences in non-traditional
designs that the workshop sought to
identify. The fourth topic area includ-
ed traditional designs that attempted
to broaden evaluation criteria to ex-
amine systems interactions.
Based on participants' responses
to the four topic areas, one or two

The FSSP newsletter is published quarterly by the Farming Systems Support Project (FSSP), which is funded by AID Contract No. DAN-4099-A-00-
2083-00 and administered by the Institute of Food and Agricultural Sciences (IFAS), University of Florida, Gainesville, Fla. 32611. IFAS is an Equal
Employment Opportunity Affirmative Action Employer. The FSSP newsletter encourages the contribution of stories, pictures and ideas, which should be
sent to FSSP Editor, 3028 McCarty Hall, University of Florida, Gainesville, FL 32611.

"key informants" with experience
in the respective topic area were
assigned to each group. Each group
also had a facilitator. Within each
group, each "key informant" des-
cribed their experiment and how it
sought to address the topic area. At
the end of the group discussion
period, the facilitator from each of
the four groups gave a brief summary
of the discussions in each respective
The second session (October 6)
focused on two key attributes of
successful non-traditional on-farm
1. Role of the farm household.
2. Assessment of systems interac-
To provide focus for the discus-
sion, only examples of adaptive (or
refinement) trials were sought. These
are trials which seek to develop
specifics for a technological change
already identified as a potential solu-
tion to the priority farmer problem.
Refinement trials thus fall between
exploratory trials, which seek to
identify a potential solution for a
wide range of technological possib-
ilities, and validation trials, which
assess farm household acceptability
of a specific technology developed
in earlier refinement trials (Caldwell
and Walecka, 1986; Hildebrand and
Poey, 1985).
"Key informants" were sought on
a volunteer basis from the partici-
pants as a whole. Participants were
divided into four groups, each with
"key informants." In each group, the
"key informants" were asked to ex-
plain their experiences in detail fol-
lowing four steps: objectives, design,
implementation, and analysis. Each
of these steps was examined in terms
of the above-mentioned two key
attributes. At the end of the group
discussion period, each group's facil-
itator again summarized the group's
discussion to the participants as a
Independent Farmer Experimentation
The two examples given by infor-
mants were of farmer experimenta-
tion stimulated by outside interven-

tions. No examples of experiments
initiated spontaneously and indepen-
dently by farmers emerged in the
workshop. The two examples were
largely used as a basis for planning
new researcher-designed experiments.
An illustrative example is the Inter-
national Livestock Center for Africa
(ILCA) alley cropping experiments
in Nigeria (Atta-Krah and Francis,
1986). Alley cropping was introduc-
ed as an improved or "rationalized"
alternative to the farmer practice of
bush fallowing. The main difference
with bush fallowing is the planned
planting of fast growing trees of eco-
nomic value, like Leueaena species.
Alley cropping is inherently a very
flexible system that can be used to
meet different objectives: leaves
from the alley-cropped trees can be
used either as livestock feed or as a
soil amendment, and trees can be
cut either for firewood or timber.
Both firewood and timber can be

used by the farm household itself,
or sold.
After describing the various poten-
tial benefits of alley cropping to
farmers, the research team let farmers
use the system as they saw fit. The
result was many farmer changes:
1. Introduction of burning (a hybrid
of bush fallow and alley cropping).
2. Farmer-develped techniques for
incorporating the foliage into the
3. Introduction of tractors.
4. Moving livestock from a free-
ranging state to confinement with
cut-and-carry feeding.
The research team used several
methods to evaluate the farmer-
introduced changes:
1. Monitoring the inputs used by
2. Farmers' own evaluation.
3. An adoption study to determine
which farmer-introduced changes
were being more widely adopted.

Table la. Matrix of information for description of objectives and design of
a non-traditional experiment in terms of the role of farmers and
assessment of systems interactions.

Steps in
experimental Role Assessment of
process household systems interactions
Objectives Based on a farmer problem- The nature of the problem com-
complex of soil, vegetation, plex forced the experimental
labor, and livestock feeding objectives to encompass several
identified through interaction components of systems. The
with farmers in the diagnostic objectives were:
survey. -to speed up soil fertility
recovery during the fallow
-to eradicate undesirable
weed growth (Cogon:
Imperta cylindrica).
-to reduce labor in cultiva-
tion at the end of the
fallow period.
-to improve pasture quality.

Design Farmers selected plot location Measurement instruments
and size within a range were developed for soil
determined by researchers. fertility, labor, pasture, and
Treatment structure and weed vegetation.
replication was determined by
voluntary farmer involvement.
Researchers determined a
critical level of observations
for each treatment level at
different points in time after
treatment implement was
done using a dynamic rolling

One difficulty is that the lack of
standardization in farmer experi-
mentation has made it impossible to
do quantitative analysis of the above
evaluation data.
These on-farm experiments helped
researchers discover areas for needed
research. The farmer-introduced
changes became the basis for choice
of priorities and factors for the re-
search team to study in more con-
trolled experiments. In essence, a
farmer-managed trial was used as an
exploratory trial, in place of more
traditional researcher-managed ex-
ploratory trials, but with the same
objective of identifying factors and
treatments for subsequent refinement
Flexible designs that respond to
farmer initiatives
From six examples given by infor-
mants and other participants, the
workshop identified four common
areas of flexibility in design:

1. Giving farmers choice over the
levels of non-experimental vari-
ables (experimental constants),
such as planting dates, spacing,
and weeding.
2. Changing treatments during estab-
lishment of the experiments on
the farm, either by changing treat-
ment levels from those in the
"paper" design, or by adding a
range of farmers' individual treat-
ments that vary from farm to farm
to a core of common treatments
placed on all farms.
3. Changing designs between seasons
so that the second year's treat-
ments reflect farmers' responses
to the first year's treatments.
4. Making changes during the course
of the experiments (after establish-
Participants also identified four
common elements and problems:
1. Analysis becomes more continual
in nature, with weekly or monthly

analysis of measurements taken as
the crop grows and the experiment
2. Flexibility in design introduces
problems of confounding and
higher coefficients of variation
(CV) in analysis. Increasing the
number of farmers is one way to
compensate for this, but that in
turn increases logistical difficulties.
The research team needs to judge
carefully the number of observa-
tions needed to test hypotheses.
3. Changes in statistical analysis resu It
from more flexible designs. Analy-
sis is more often based on unbal-
anced data. The focus of analysis
changes from means probabilities
to robustness, for example through
confidence intervals.
4. Making changes in treatments dur-
ing the course of the experiment
(after establishment) is easier for
long-term experiments with inter-
ventions spread over several sea-
sons, such as perennial crops and
soil conservation techniques, than
with annual crops. Over several
seasons, the experiment can build
up a body of data with enough
observations to test hypotheses.
The example that best illustrated
change during the course of the ex-
periments was an example from the
Philippines (Lightfoot et al., 1986).
This experiment is testing legumes
planted in place of traditional un-
planted fallow vegetation. The ex-
periment also illustrates the format
used in the second session for des-
cribing non-traditional experiments
by examining objectives, design, im-
plementation, and analysis in terms
of the role of the farm household
and assessment of systems interac-
tions (Tables la and 1b).
Analysis of Systems Interactions
Systems interactions are wide-
spread, since crops and animals
usually have multi-purpose uses. Yet,
people often do not see these even in
their own experiments. For example,
in one informant's experiment, plow-
ing after rice harvest prior to planting
maize was introduced in a trial as a
means of reducing weeds and increas-
ing maize yield. However, the most

Table 1b. Matrix of information for description of implementation and an-
alysis of a non-traditional experiment in terms of the role of
farmers and assessment of systems interactions.

Steps in
experimental Role of Assessment of
process household systems interactions
Implementation Farmers established treatments. Researchers monitored plots
Researchers provided legume for soil fertilization, labor
seed. In some cases, researchers use in establishments, legume
also helped sow, although this competition with Cogon, and
researchers input was contrary pasture quality. Researchers
to planned design. Farmers also obtained an overall farm
implemented the grazing household assessment of the
regime and determined when experiment.
when to cultivate.
Analysis Comparison of farmer responses Data from each component
and quantitative measurements, was analyzed as the data came
For example, farmers said in (for example, labor and
establishments was too laborious legume analysis shown in the
and measurements of hours paper). Up to this point, anlaysis
worked confirmed this. Growth has been mainly for interpreta-
data showed slow percentage tion. Standard deviations show
of cover, the range of variability of each
treatment and the degree of
overlap of treatment results at
each point in time. More sophis-
ticated predictive analysis of
differences and trade-offs among
effects on different systems
components of the treatments
will become possible as more
observations added at each
time point.

limiting resource of the farm house-
holds was not maize yield, but rather
draft power. Hence, farm households
preferred to relay-plant maize into
the rice fields without plowing, even
at the expense of maize yield, in
order to save draft power.
Examples from Indonesia (Pierce
Colfer, 1986) and Mali (Verbeek et
al., 1986) indicated how concurrent
collection of socio-economic data on
labor use, tillage methods, and animal
use brought more farm household
criteria into trial evaluation. Such
evaluation using more systems inter-
actions criteria in turn helped fine-
tune treatments and even redirect
trial choices.
On the other hand, all informants
had difficulty combining soci-eco-
nomic survey data and plot experi-
mental data in a consistent, analyz-
able format. For example, in Nigeria
(Palada et al., 1986) a maize-cassava
intercropping trial included farmer
criteria of taste and storage, and
developed an index which combined
the yields of the maize and cassava.
However, the trial did not produce
examples of how to assess system
trade-offs among yield, home con-
sumption, sale, taste, and storability.
The workshop uncovered fewer
examples of independent farmer
experimentation, flexibility in de-
sign, or analysis of systems inter-
actions than anticipated. The work-
shop was also handicapped by not
having asked informants to bring
actual data sets. Nevertheless, the
workshop was valuable in indicat-
ing some of the problems of the
current "state-of-the-arts" in the
The workshop resulted in three
sets of recommendations:
1. More examples of data based on
independent farmer experimenta-
tion, data from unbalanced designs
resulting from farmer modifica-
tions, and data combining measure-
ments on different components of
the farming system (biological and
socio-economic) are needed. Read-
ers with data sets falling in any or

all of these three categories are
requested to return the enclosed
form or write to FSSP. This infor-
mation will be used in planning for
the 1987 FSR/E symposium.
2. A session to which participants can
bring "problem" data sets should
be included in the 1987 sympo-
sium. Such a session should be
planned and announced well in ad-
vance of the symposium. It should
last at least a half a day, be part of
the regular symposium (rather than
a pre-workshop session), and in-
clude an opportunity to report on
results in the final closing plenary
session. Persons involved in FSR/E
training and training materials de-
velopent should be invited to par-
ticipate in the workshop. A statis-
tical consultant familiar with the
objectives of FSR/E and the
reasons for the need for new
techniques of design and analysis
should also be available during
the workshop.
3. Training manuals and workshops
should include the following
a. The importance of seeing as
many farmer plots or animals
as possible during diagnosis, to
identify independent farmer
experimentation. An example
is given in a recent paper by
Ocado et al. (1986).
b. Additional techniques of analy-
sis, with examples, for unbal-
anced data for which an F-test
may be impossible or inconclu-
sive. Such techniques include
scatter diagrams, standard errors
of treatment means, and confi-
dence intervals.
c. Procedures, with examples, for
identification of systems inter-
A limited quantity of a more
detailed report of the workshop is
also available upon request to
the FSSP.N
Alta-Krah, A. N., and P. A. Francis, 1986.
The role of on-farm trials in the evalua-
tion of alley farming. Paper prepared for
the International Workshop on Alley
Farming for Humid and Sub-humid
Regions of Tropical Africa. I LCA Humid

Zone Program, Ibadan, Nigeria (mimeo).
Barker, R., and C. Lightfoot. Farm experi-
ments on trial, pp. 300-321. In: C. B.
Flora and M. Tomacek (eds.). Farming
Systems Research and Extension: Man-
agement and Methodology. Paper No.11,
Kansas State University, Manhattan,
Caldwell, J. S. (technical editor), and L.
Walecka, (coordinating editor). 1986.
Techniques for design and analysis of
on-farm experimentation. FSR/E train-
ing units: volume II. Farming Systems
Support Project, Gainesville, Florida.
Hildebrand, P. E., and F. Poey, 1985. On-
farm agronomictrials in farming systems
research and extension. Rienner Pub-
lishers, Boulder, Colorado.
Lightfoot, C., T. Cornick, R. Ayaso, Z. de
la Rosa, E. Lapasanda, G. Aves, and P.
Hipe. 1986. A short methodological
account of a dynamic systems field
experiment: the case of legume enriched
fallows for the restoration of soil fertil-
ity, eradication of Imperata, improve-
ment of pasture, and reduction in labor
for cultivation, in the Philippines. Paper
presented at the 1986 Farming Systems
Symposium, Kansas State University.
October 5-8, 1986. Farming Systems
Development Project-Eastern Visayas,
Ministry of Agriculture and Food, Region
VIII, Visayas State College of Agricul-
ture, and Cornell University, Tacloban,
Leyte, Philippines. (mimeo).
Ocado, F. D., A. D. Jumaday, A. L. Aliman,
and C. L. Lightfoot, 1986. A study of
farmer's adaptation of upland rice to
rainfed bunded conditions. Farming
Systems Development Project-Eastern
Visayas Working Paper No. 3. Ministry
of Agriculture and Food, Region VIII,
Visayas State College of Agriculture,
and Cornell University, Tacloban, Leyte,
Palada, M. C., W. O. Vogel, and H. J.
Mutsaers. 1986. On-farm testing of
improved technologies in Southwestern
Nigeria: the IITA experience, pp. 383-
419. In: C. B. Flora and M. Tomacek
(eds.) Farming Systems Research and
Extension: Food and Feed. Paper No.
13, Kansas State University, Manhattan,
Pierce, Colfer, C. J. 1986. Social science
contribution to soil management: The
Tropsoils example, from the view of an
anthropologist. Paper presented at a
symposium at the University of Hawaii,
June 23, 1986. International Agricul-
tural Programs Office, University of
Hawaii, Honolulu, Hawaii (mimeo).
Verbeek, K., B. Sanago, and P. Kleene.
1986. The FSR&D-extension linkage:
experience from Mali, pp. 152-164. In
C. B. Flora and M. Tomacek (eds.)
Farming Systems Research and Exten-
sion: Food and Feed. Paper No. 13,
Kansas State University, Manhattan,

Building ORF into Existing Agricultural

Research and Extension Institutions*

Presently in the Eastern and South-
ern African region research and
extension are divided institutionally,
often within the same ministry but
relatively autonomous. Present re-
search organization usually takes the
form of research stations with multi-
disciplinary commodity or specialist
research teams with national respon-
sibilities for their specialization. In
larger countries there may be many
such stations with the commodity or
specialist teams at each sometimes
focused on commodity or disciplin-
ary issues of particular importance in
their immediate region. Within this
historical pattern there are at least
four options for the organization of
systems based OFR. Each option has
advantages and disadvantages in terms
of its ability to perform the roles
designated for systems based OFR
and in terms of the ease of institu-
tionalization and arrangements within
existing research and extension insti-

Each Commodity/Specialist team
in NARS adds a social scientist to
bring the farming systems perspective
into their research. All scientists of
the team carry out both technical
and on-farm research related to the
team's commodity or specialization
across its area of responsibility i.e.
nation or region wide.
Little institutional change or

management re-organization is
required for implementation.
* The technical/or farm research
interaction is inherent, within
the team.
* Researchers keep a foot in 'real'
research with promotional pro-
spects and peer recognition on
traditional, established criteria.
They feel less threatened than
being transferred to a 'new'
type of research in which novel
criteria are relevant.
Disadvantages and difficulties
* The pre-determined focus onto
the team commodity or special-
ization limits the benefits from
the systems perspectives.
-research effort may be focus-
sed on what to farmers are
relatively minor problems.
Identified solutions may not be
attractive to farmers whose
overriding concerns are else-
where in their system.
-technical research problems re-
quiring CST attention can only
be identified and ranked within
the special focus. This option
cannot aid prioritization of CST
efforts across commodities and
specializations-a major contri-
bution from a full systems per-
* There is potentially massive
overlap in the farmer focus of
CS Team based OFR. (At the
extreme five or six teams may
be carrying out a special focus
diagnostic and OFR experimen-
tal program within the same
target group of farmers).
* The disadvantages listed for this
option contribute to the limited
cost-effectiveness of systems

based OFR when integrated
with CS Teams.
The linkage role with extension
is inhibited in this option. Little
re-organization is required and
OFR's primary concern is the
commodity or specialization.
This is relatively incompatible
with area oriented extension
Unless the social scientist added
to the CS Teams is very experi-
enced in systems based OFR it
is very difficult for him to make
the case for social science and
carry the team of technical
scientists with him in introduc-
ing a systems perspective. Such
experience is o far rare in

Each Commodity/Specialist Team
in NARS has a complementary OFR
team (usually an agronomist and a
social scientist) which handles the
systems based OFR work relevant to
the CST throughout its area of respon-
sibility i.e. nation or region wide.
Advantages and Disadvantages
This has similar disadvantages
to Option 1 with respect to the
restricted exploitation of the
systems perspective both in
mobilizing technical research
results and in drawing priorities
across CST's. It requires more
re-organization than having all
CST scientists involved in both
technical and on-farm research.

*Farming Systems Newsletter, July-September 1985, CIMMYT Eastern African Economics Program, International Maize and Wheat Experiment Centre
(CIMMYT). P.O. Box 25171, Nairobi, Kenya.

* It also asks OFR scientists to
desert the traditional peer
groups evaluation criteria in
research and face the lack of
an OFR career structure in
many NARS, risking uncertain
promotional prospects.
* It provides a clear institutional
niche for scientists, allowing for
development of their capacity
in OFR before being drawn into
'confrontation' with technical

The set of Commodity and Spec-
ialist Teams are complemented by a
set of systems oriented OFR teams
in NARS. Each OFR team has region-
al responsibilities, each draws from
and feeds back to all or any CST's
which are relevant to circumstances
and priorities of farmers in its region
of responsibility.
In diagnosing focii in the sys-
tem without pre-determination
to a particular Commodity or
Specialist, priority problems are
identified. It is these priorities
which offer best leverage for
improvement of the system,
and appropriate solutions for
these should be readily absorbed
by target group farmers.
Feeding back technical research
agenda to CST's made up of un-
solved priority farmer problems
helps balance technical research
efforts according to identified
farmer needs.
The two points above and the
responsibility of one OFR team
for any one region make a strong
contribution to a cost-effective
research effort.
The regional or area orientation
of theOFR teams is wholly com-
patible with extension organiza-
tion. It offers great potential for
drawing extension staff into the

later stages of generation of the
techniques they will later have
to sell to their farmer clients.
It helps extension staff identify
with technologies incorporated
into recommendations.
* It creates an institutional niche
for scientists where they can be
sheltered during orientation,
while their professional compe-
tence in OFR is built up.
* The separation from CST's may
have several adverse effects:
-It asks scientists to isolate them-
selves from traditional peer
groups criteria in assessing re-
search programs.
-It may cause uncertainties as to
career structure and promotion-
al opportunities.
-Draw down from and feeback
to CST's are indispensable fea-
tures of the complementary

With each country a unique situa-
tion there is no universally best
option for incorporating a systems
based on farm research approach in
national research and extension
services. The options must be weighed
by decision makers in the light of
their specific country circumstances.
The variability of farming situa-
tions across the country-the
more variable the more local
specific will be technology
The complexity of farming sys-
tems-the more complex, the
greater the potential contribu-
tion of a systems perspective in
research prioritization and the
generation of appropriate tech-
The existing organization of re-
search and attitudes of research-
ers. It may be important to min-
imize re-organization. It may be
counter productive to risk alien-
ation of the existing research
The existing linkage between
research and extension-where

roles of CST's and OFR teams.
By separating the two sets,
albeit within NARS, there is a
danger of poor linkages be-
tweeen them.
0 A more complex re-organization
of institutions and budets is re-
quired to implement this option.

This is the same structure as option
3 but with the OFR teams institu-
tionalized within the extension ser-
vices, not in NARS. It has the same
advantages as option 3. It has the
added disadvantage of the complete
isolation of the CST's and thus greater
dependence on feedback loops for
the effective application of the sys-
tem perspective by the OFR teams.

this is effective then less weight
need be given to reconciling the
structure of OF R with extension.
Analogous situations with strong
technical research efforts in
other countries of the region-
this will, given regional cooper-
ation, influence the balance
between technical research by
CST's and OFR in country.
The size and complexity of the
country affects the desirable balance
between technical and on-farm re-
search and the choice of OF R option.
At the extreme, very small countries
with larger, agro-ecologically similar
and cooperative neighbors or near
neighbors, may choose OFR and rely
on cooperation to provide the tech-
nical stockpile for domestic OFR to
utilize. Larger countries may do this
for areas which have analogous
conditions elsewhere in the region
with strong technical research efforts.
There is massive scope for regional
cooperation between countries in
agricultural research in which the
need for a critical mass of researchers
is recognized but domestic resources
-in terms of both manpower and
budgets are often limited.i

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