10.3 Agronomy and FSR -A Reluctant Marriage ? Peter Hildebrand and Dennis Keeney.
Within agronomy there have always been important systems related themes. Rotational experiments
are perhaps the oldest example. Since the 1930s there have been forays into intercropping to improve small
farmer husbandry, particularly in Africa in the 1950s (eg Evans 1960). As was seen in Dick Harwood's
contribution in chapter two, the cropping systems initiatives of Richard Bradfield evolved into another important
system related theme. Bradfield was also one of the first agronomists to recognize the value of working
cooperatively with economists. (Bradfield 1966)
Since the mid-1960s the involvement of agronomists in FSR has added a further systems theme,
perhaps with greater repercussions for the discipline than earlier ones. It has brought agronomists closer to
their resource poor clientele in developing countries operating complex farming systems. It has brought
experiments off the research station, out of a controlled environment, onto farmers' fields to face the vagaries
of nature and be managed by local farmer practice. While it would be wrong to say that agronomy has
changed as a result, it would be fair to say that as a result of FSR a significant number of agronomists are now
working in an interdisciplinary way and using non-traditional methods for designing and evaluating technology.
As earlier chapters have shown, FSR was born from professional disquiet at the low acceptance of
research recommendations by small farmers. Yet the accumulated wealth of conventional agronomic skills and
methods were the very vehicle which launched FSR into the development of more appropriate technologies.
Conventional agronomy was the foundation of this key application of FSR. The more difficult question, and the
one which occupies the remainder of this contribution, was how has the application of FSR to technology
development influenced agronomy itself.
Five main sources of impact on agronomy are identified. All five arise from the routines of FSR; working
with small farmers on their fields, working with an awareness of the farm as a system, and interacting with social
scientists. The five sources of impact are;
The design and evaluation of technologies in farmers' fields.
Appreciation of the variable quality of the land resource small farmers manage.
Understanding the interactions between one enterprise and the rest of the farm
Understanding the diversity among small farmers.
New priorities for formal, station-based, agronomy.
Before examining each of these sources, the interaction with social scientists deserves
10.3.2 Working with social scientists.
Chapter two suggests the introduction of social scientists into agricultural research in
developing countries as an important origin of FSR. It sprang from criticism, implicit but
often also explicit, of the relevance of the products of the existing establishment. The
'hard' scientists in the research services, with agronomists and breeders at the centre,
saw these young newcomers (they were inevitably junior professionals) as 'new boys on
the block'. In establishment terms that meant being seen but not heard. Not only were they
seen as new, but also as representatives of a soft science with little apparent relevance
to the conventional paradigms which had dominated agricultural research for the last fifty
years. Yet these junior irrelevancies, almost before they had their feet dirty, were criticising the way in which
research was being organised and managed! Understandably there was early resentment against the
introduction, often seen as intrusion, of social science into agricultural research.
10 December 1997
Perhaps the first break was the ability of economists to use money as a common denominator, and
interpret hard science results in a way familiar to scientists themselves, from their personal housekeeping, in
terms of budgets. There was increasing appreciation that economic evaluation added to relevance. In this
early mode there was little threat; the economist remained dependent on the scientist for data, and was seen
in a service role which was acceptable. However, sometimes even that role had to record economic failure for
what had seemed a biological and statistical triumph, and wounds were opened. For example, in the
interpretation of a fertilizer response curve: While the agronomist concentrated on levels realising the highest
yields, the social scientist needed to nett out the costs of fertilizer and the costs of applying it. Yet cash is so
scarce among resource poor farmers that highest money return per unit of cash outlay was a much more
appropriate criterion, but always much lower down the curve than the technical optimum. Wounds deepened
when social scientists began to agitate about experimental design, and, working with farmers, to encroach into
the territory of agronomists and breeders by putting down their own experiments. Common comments from
agronomists about the bedraggled nature of FSRE trials under real, limited-resource conditions were
something like, "This looks just like a trial being run by social scientists" to, "It's a good thing it is well off the
road!" One of the most usual and most disheartening comments was, "It is obviously not worthwhile to work
under these conditions because nothing can be accomplished" (Hildebrand 1979). However it was in these
circumstances, when agronomists themselves were drawn out onto farms to interact with farmers and to learn
about their production conditions that social science skills demonstrated their value. In eliciting and
understanding farmers' goals, priorities, strategies and constraints, social scientists clearly could provide
information for the design and evaluation of more relevant experiments.
It is still too early to say that the battle is over and the disciplines are reconciled, there are still too few
research establishments with a social science cadre. Today however, there are many examples of powerful
partnerships between farmers, agronomists and social scientists.
10.3.3. The design and evaluation of experiments in farmers' fields.
The move from the artificial environment of the station, and from the rigid ANOVA conventions, into the
real world, has been a real challenge to the agronomy profession.
The move off the station. The research station gave agronomy a safe haven from the myriad sources of
variation designed to frustrate the measurement of responses. Fertility was managed, ad lib water was often
provided, and machinery allowed precise control of timing and consistency in method. Even where labour was
used on stations it was inevitably provided ad lib under the banner of precision. Such facilities did indeed
provide an environment in which agronomists and breeders could work with precision. However, not only is
such a context insulated from the vagaries of the real world, it also has the potential to be wholly misleading.
One only has to make the facile comparison of the depth and speed of cultivation with a heavy tractor
compared to a hand hoe to see that results on research stations may not be capable of replication in
Fostered by FSR, many agronomists moved out onto those fields. This was by no means the first sally off
the station. Multi-locational trials to capture the effects of wide variations in climate, soil, and pest and disease
complexes were common. However, the locations chosen for these were treated as mini stations to ensure
the effects of spatial differences in climate, soils and pests could be isolated and accurately measured. Such
trials, while playing an important role, are a poor basis for recommendations to farmers. The move off-station
prompted by FSR was different and had a range of repercussions. Many of these caused by increasing
familiarity with small farmers as partners, and with the conditions which such farmers must manage.
System interactions. As early as 1980 Henry Nix, an Australian agronomist, was proactively urging a wider
systems perspective for the profession;
"A research strategy based on the systems approach would centre around the development of working
models of crop production systems. Such models need to be structured so that they remain operational yet
capable of continuous improvement in logical structure and function. Ideally it would be useful to have a
hierarchy of models capable of application at a range of scales and offering some choice in the levels of
precision and accuracy." (Nix 1980)
Intercropping and cropping systems research widened the range of interactions to be investigated.
Farming systems research took this a step further and superimposed the whole set of socio-economic
constraints for consideration while shaping agronomic improvements for farmers. The interactions between
enterprises in their demands for cash, labour and indeed land, and the complementarity of enterprises,
particularly livestock with crops or crop residues for feed, manure and draught power for crops, brought a new
dimension to agronomy. Even the fact that increased cotton production is a national priority will not make
farmers plant cotton early, and use agronomically perfect husbandry, if their own priority is for an early
planted, early available food crop to fend off starvation. In such circumstances farmers must plant their food
crops first, then plant their cotton. Analysing such an interaction tells us that greater efficiency in early planting
of food crops, or cheap supplies of food in the local market in the hungry months, are alternative routes
towards better husbandry in the cotton crop.
System interactions, an understanding of how farmers use these and put priority on some interactions,
widen both the experimental hypotheses and the evaluation needs They also imply the importance of
compromise solutions which, while sacrificing the optimum from any one enterprise, enhance overall system
Within farm and within field variation in soil and water resources. The exposure in farmers' fields also led
agronomists to an appreciation of the variation there, and an understanding that farmers used this in managing
their resources (see Lightfoot, section 12.6, this volume); for example in exploiting local differences in
waterholding capacity in different parts of a single field. Strong interactions were apparent between key
management factors and such resource 'niches' that farmers were using in a particular way. In rainfed
farming the need for 'niche' technologies, rather than broad adaptability began to emerge.
Farmer Diversity. Exposure to farmers, the concerns of the social scientists, and the need to identify
representative clients for partnership in on farm experiments, all exposed agronomists to the diversity of the
farm population. They learned that old farmers have a different attitude and different capabilities to young ones,
and women headed households usually have a weaker labour force than a full family. This diversity brought
further evidence of the futility of a single best answer. It reinforced the need for baskets of choices, tailored
both to the diversity of households in local communities, and to the variety of resource niches used by them.
Experimental design. Much of the influence of FSR on agronomy as a discipline has arisen from adapting
conventional methods, particularly the analysis of variance, for use in the real conditions which small farmers
must manage in order to produce. The classic agronomic goal of producing "broadly adaptable" technology
fit in well with the practice of controlling non-experimental variables at non-limiting levels. This practice created
artificially superior environments that large-scale and/or industrialized farmers could mimic but limited-resource
farmers could not achieve (Hildebrand and Russell 1996).
Conventional agronomy identified treatment variables as those management factors most likely to
contribute to high yields per unit area under relatively homogenous conditions of climate and soil. Experiments
sought optimum levels in each of these key factors. For precise measurement of the responses to changes
in their levels, non-experimental variables were, and are, conventionally held at high, non-limiting levels. As
with the false environment of the station this completely isolated the results from small farmer management.
Recommendations are normally framed from the findings on the treatment variables alone implicitly assuming
that non-treatment variables would be non-limiting when such recommendations were implemented on the
fields of small farmers. This was never the case. Small farmers never had the machinery, rarely had the cash,
and often hadn't the labour, to implement the treatment recommendations, let alone the resources to
implement the non-treatment management at non-limiting levels.
Overall it was small wonder that there was a lack of adoption, few recommendations were accepted,
particularly in dryland farming where the uncertainties of climate also had to be managed. In FSR driven on-
farm research, farmer practice became the standard for non-treatment management. Responses obtained to
treatment factors were much closer to those to be expected when treatments were tested by farmers.
In the early days there was a strong reaction from agronomists to this loss of control, and to an extent the
reaction continues. By their training agronomists inevitably associated the relatively low yield levels and the
limited treatment effects with high cvs and lost trials. Reconciling statistical rigour and reliability with relevancy
in experimentation has been a real area of conflict for agronomists in FSR, perhaps it is the heart of the
'reluctant marriage' label. In much on farm research the balance has shifted to relevancy, with replicability
more a function of significant numbers of farmers testing out apparently relevant new materials and methods
on a 'suck it and see' basis. This has been supplemented with agronomists thinking more holistically about the
circumstances of the season to understand the conditions that realized a particular experimental outcome, and
examining these in an across season context, and asking wether it will happen often enough to be a regular
success? When the ANOVA is used at appropriate points in the process, novel designs have helped to
reconcile highly variable field conditions, logistical dilemmas, and the conventional analysis; for example the
use of farms as replications and the wider use of gradients.
Similar reservations have arisen about conventional breeding strategies. It has long been argued that
conventional selection methods have had a negative effect on the relevance of germplasm made available to
small famers (Republic of Zambia 1983). Hildebrand (1990) lists four factors likely to have caused the rejection
of genetic material which would have demonstrated superior yielding abilities in both the poorest and the best
Statistical dependence on ANOVA that leads to the concern with reducing genotype by environment
interactions, which in turn leads to;
The nearly universal practice of evaluating material on experiment stations and farms with real or
artificially created superior environments to control this interaction or to permit the material to express
its yield potential
The capability of many farmers in the developed world, over the last few decades to use their
resources to modify unfavourable environments; and
Widespread use of a regression coefficient of unity as a measure of stability.
With FSR at least partially responsible, the recent surge in participatory breeding practice has
overtaken the theoretical discussion. The widening understanding of the farming system, the fact of
niche management, and of small farmers' use of diversity as a management tool, has brought new
selection criteria to bear through direct farmer involvement in varietal choice (Sperling et al. 1994, and
section 12.5 this volume). The burgeoning interest in diversity as an environmental goal has helped
spur efforts to service farmers' needs for a range of plant material.
Experimental evaluation. Working with farmers in their fields raised awareness among agronomists
of the real circumstances under which small farmers operate. They became more familiar with the
choices small farmers must make, and the criteria they used in making these. Yield per unit area was
never their basis for choice. Return per unit of cash or labour outlay were key criteria for resource poor
farmer. Farmers' criteria led to the need to modify conventions in evaluation and eventually in the
choice of experimental treatments. However, given the priority for most small farmers to meet their
household food needs day in day out, and given the need to manage risks from uncertain rainfall and
uncertain markets and prices, the trade-offs across multiple objectives and a range of evaluation
criteria become complex and subtle. So much so that farmers themselves have proved to be the best
evaluators, and many FSR teams now involve them as such. These factors, and a need to shift
ownership of on farm trials to the farmers and their communities, have led to wide farmer participation
in trial evaluation, and increasingly in experimental design.
Setting on station priorities. Finally, agronomic priorities for controlled station research have begun
to respond to information from diagnosis and from the closer association of agronomists with both
farmers and social scientists. Station trials increasingly investigate the underlying biophysical
relationships of the problems thrown up by farming systems research on farmers' fields. It is a start to
the articulation of demands from the hitherto unorganised small farmer population and is influencing
priorities in both breeding and agronomy.
10.3.4 Effects in Professional Circles.
Within agronomy a dichotomy seems to be developing; some, essentially field agronomists,
are now devoted to OFR. For others the rise of the sustainability issue has either made them more
reductionist in approach, focussing narrowly on the detailed physical processes governing the
relationships between soils, water and plants, or has pushed them higher up the systems heirachy to
the broader fields of ecology and soil and water processes at the watershed level. In some senses the
battle for wider systems thinking among agronomists has been won. A significant sub-set are
broadening out into farm level systems (as opposed to plant growth systems) and have found a
professional home in the FSR associations, particularly in Africa, Asia and Latin America. Beyond this
systems agronomy is also well represented in conventional professional circles, certainly in the USA.
As early as 1976 the American Society of Agronomy (ASA, CSSA, SSSA) published "...
the first major publication of the three societies on one aspect of increasing food production beyond
increasing cultivated area and increasing yields; that is, harvesting more than one crop from the same
piece of land in a year." (Papendick et al. 1976, p. v). This publication resulted from an ASA
symposium led by the International Agronomy Division (A-6), on multiple cropping at which two social
scientists (both agricultural economists) participated. In 1978, a symposium, led by the Extension (A-4)
and International Agronomy Divisions, on transferring technology for small-scale farming had a
participating anthropologist and resulted in another ASA publication. In this book, then ASA President
John Pesek (1981) argued the importance of small farmers as providers to the world supply of food,
and Nyle Brady, then Director General of IRRI, lamented that only about one fourth of the rice farmers
in the tropics had benefitted from the improved rice technologies. "For the other three quarters, no
really superior technologies have been developed that suit their conditions and that financially benefit
them" (Brady 1981, p. 7). Brady further argued that "A major constraint to modification and adoption
of new technology to small-scale LDC farmers is the failure of researchers and extension personnel
to work with them" (Brady 1981, p. 9). In the same publication, Robert Waugh, Adjunct Director of
ICTA in Guatemala, argued that "Probably the use of multidisciplinary teams of biologists and social
scientists has contributed more than any other thing to making it possible for the social scientists [at
ICTA] to contribute to agronomic technology. By the same token, this integration has had a beneficial
effect on the nature of the work undertaken by the biological scientists" (Waugh 1981, p. 85).
Many members of the Inernational Agronomy Division have long been active in FSRE, and the
division elected an FSRE agricultural economist as Chair Elect in 1990. In 1994 the ASA created the
Agricultural Systems Division (A-8) with support from the FSRE community within and outside the ASA.
In the December, 1995, Agronomy News, the ASA announced that its premier publication, the
Agronomy Journal, would open a new topic section, Integrated Agricultural Systems, which would
include farming systems research and extension.
10.3.5 The Ongoing Honeymoon.
Continuing interactions between biophysical and social scientists, many within the context of
FSR and FPR, are the honeymoon following the reluctant marriage. There are recent developments,
particularly in the areas of natural resource management and participatory breeding.
Agronomists have increasingly associated FSR with sustainable agriculture. Many
agronomists with FSR backgrounds and experience in developing countries have been at the forefront
of sustainable agriculture conceptualisation in the USA, undoubtedly because of the systems
approach underpinning FSR and the location-specific nature of technologies required by limited
resource farmers. Obviously, FSR has not been the only influence on the move toward systems
approaches to sustainable agricultural development, but its influence is undeniable. When the USDA,
National Research Initiative, Competitive Grants Program initiated its new Agricultural Systems Division,
an FSR agricultural economist headed the proposal evaluation team that had several FSRE agronomy
professionals on it. The USAID-funded Sustainable Agriculture and Natural Resource Management
(SANREM) Collaborative Research Support Program (CRSP), the USDA-funded Sustainable
Agriculture Research and Extension (SARE) program, and the reorganized Soils Management CRSP
all require methodologies that were pioneered by FSR practitioners.
Iowa State University (ISU) agronomist, and former ASA President John Pesek chaired the
Board of Agriculture's National Research Council committee that prepared the 1989 NRC book on
Alternative Agriculture. Dennis Keeney, another ISU agronomist and also a former ASA President
currently heads the Leopold Center at ISU. But as he points out, administrative declarations do not
change paradigms. After a decade of Leopold Center funding and persuasion there are only two
FSRE-type projects in Iowa that truly involve agronomists, even though there are and will continue to
be much research and education in alternative systems.
In both agronomy and breeding, at least among the cutting edge members of the disciplines,
there is acceptance that technology is not 'farmer size neutral', small farmers have particular needs.
There is accumulating evidence that farmers out perform breeders for 'niche' environments and niche
breeding with farmer participation is rapidly establishing a role, particularly in areas and crops where
market penetration is weak. There is increasing specialisation amongst agronomists as the field both
widens, to the farming system and up the hierachy to micro-ecologies and watersheds, and deepens
into the soil/water/plant relationships becoming increasingly important under the sustainability banner.
FSR can claim to have brought agronomy to terms with a set of farmer clients operating under
circumstances very different from the commercial farmers among which it grew up. At the same time
FSR must acknowledge that without the accumulated skills of conventional agronomy it would have
had no vehicle to reach out to farmers in their fields in developing countries.
Bradfield, R. 1966. Toward more and better food for the Filipino people and more income for her
farmers. ADC Paper, The Agricultural Development Council, New York.
Brady, N.C. 1981. Significance of developing and transferring technology to farmers with limited
resources. p. 1-21 In Usherwood (ed.) Transferring technology for small-scale farming. ASA Spec.
Publ. 41. ASA, CSSA, SSSA, Madison, WI.
Committee on the Role of Alternative Farming Methods in Modern Production Agriculture, Board
on Agriculture, NRC. 1989. Alternative agriculture. National Academy Press, Washington, D.C.
Evans, A. 1960. Studies of Intercropping. I, East African Agricultural and Forestry Journal, XXVI.
Hildebrand, P.E. 1979. Incorporating the social sciences into agricultural research: The formation
of a national farm systems research institute. Institute de Ciencia y Tecnolog!a Agrlcolas, Guatemala,
and The Rockefeller Foundation, New York.
Hildebrand, P.E. 1990. Modified Stability Analysis and On-Farm-Research to Breed Specific
Adaptability for Ecological Diversity. In Genotype-By-Environment Interaction and Plant Breeding, ed.
Manjit S. Kang. Louisana State Agricultural Center.
Hildebrand, P.E. and J.T. Russell. 1996. Adaptability analysis: A method for the design, analysis and
interpretation of on-farm research-extension. Iowa State University Press, Ames.
Nix Not listed
Papendick, R.I., PA. Sanchez and G.B. Triplett (ed.) 1976. Multiple cropping. ASA Special
Publication 27. ASA, CSSA, SSSA. Madison,WI.
Pesek, J. 1981. Forward In Usherwood (ed.) Transferring technology for small-scale farming. ASA
Spec. Publ. 41. ASA, CSSA, SSSA, Madison, WI.
Republic of Zambia, 1982. 'Plant Breeding for Low Input Conditions'. Proceedings of a workshop on
plant breeding for 'lousy' conditions held at Mount Makulu Central Research Station, Chilanga, Zambia.
Government Printer, Lusaka.
Sperling L., M.E. Loevinsohn and B. Ntabomvura, 1993. Rethinking the farmer's role in plant
breeding:local bean experts and on-station selection in Rwanda. Experimental Agriculture 29.
Waugh, R.K. 1981. Research and promotion of technology use. p. 67-88 In Usherwood (ed.)
Transferring technology for small-scale farming. ASA Spec. Publ. 41. ASA, CSSA, SSSA, Madison,