Tailoring Research and Extension to Support
Ecologically Sustainable Agriculture

E. Ann Clark, Plant Agriculture, University of Guelph, Guelph, ON (eaclark@uoguelph.ca)

Presented to the 11 February 99 Professional Development Program of the Northeast SARE Group

The USDA-SARE program is intended to "to increase knowledge that helps farmers adopt production and marketing practices that are profitable, environmentally sound, and beneficial to local communities and society in general." Due to constraints in time, I will not comment further on the worthy topics of marketing or rural community health, but will confine my remarks to the issues of profitability and ecological sustainability.

Profitability is a fickle goal and one that is highly vulnerable to manipulation by government policy. What is economically rational in one country may not be another, depending on the direction given by government to producers through such factors as: The same process or practice can differ in profitability among comparable enterprises. In Canada, for example, practices which are profitable in the dairy and feather industries are unprofitable in parallel livestock sectors - beef or pork - owing to government policies on supply management. Furthermore, because policies change, both at home and abroad, impacting everything from supply and demand to the value of inputs and commodities, what is profitable one year may not be in another. And finally, it is at least arguable that commodity-specific supports not infrequently create an inherent conflict between profitability and sustainability, leaving producers little choice as to where to place their allegiance (Faeth et al., 1991; Faeth, 1995). Subsidy-induced overgrazing of hill land by sheep in NZ, producing severe erosion on steeply sloping land, contrasts sharply with the proliferation of soil-saving treed plantations on the same land in NZ in the post-subsidy era.

It is therefore suggested that the goal of profitability is better addressed at the level of policy than as a research objective, with the specific goal being to enhance the complementarity of profitability and sustainability. As proud and independent as producers are, they nonetheless do what they are told when it comes to profitability. If we as a society don't like what they are doing, then we had better change the policies that are directing their decisions, instead of blaming the producers. Profitability is an essential feature of farm life, but is too fickle and malleable to be intertwined with something as fixed and unchangeable as ecological sustainability (Clark and Weise, 1993).

Ecological Sustainability. It is proposed that the ecological sustainability of agriculture is founded on immutable principles, which can be transformed into a range of diverse, applied, context-specific practices. Hypothetical examples of such principles might be:
  1. Practices and processes which mimic nature are most likely to approach sustainability (e.g. minimal soil disturbance rather than inversion of the plow layer). 
  2. Efficient nutrient cycling is promoted by a cropping system characterized by net nutrient sinkness throughout the mineralization season, as via perennial ground covers or the combination of annuals and cover crops.
  3. Crop water use efficiency is promoted by practices which promote infiltration, at the expense of runoff, and transpiration instead of evaporation.
  4. Strategic biodiversity will maintain a balance among species, keeping insect and disease populations from becoming pestiferous.

I would not presume to define the key principles here, a task which has been undertaken by others, but would suggest that lack of a consensus in identifying these guiding principles must surely constrain progress toward sustainability. If empirical evidence is not sufficient to achieve general agreement on the principles, then let them be established through a standard process of hypothesis testing. But one way or the other, a critical starting point for genuine progress towards ecological sustainability must be consensus on "first principles".

I will state with conviction, however, that identifying and then tailoring the principles of ecological sustainability to commercial agriculture will demand approaches which are fundamentally different from those which fuelled resource-intensive agriculture in the past half century. When the goal is not yield but ecological sustainability, research and extension will differ not just in the questions asked, but also in how we ask them. The differences in "questions asked" are probably self-evident, but to clarify why "how we ask them" must also differ, let us consider some of the assumptions which underlie contemporary ag research.

At least as practiced in recent decades, agricultural research and extension have been predicated on unstated and perhaps even untested assumptions which may not hold when applied to sustainable agriculture (SA). These assumptions have defined the design and conduct of field research, and hence, the beneficiaries of the research. .I'd like to challenge to you reconsider and perhaps re-test these assumptions, with specific reference to training both researchers and extensionists mandated to serve the needs of SA. .
Query: Has the validity of this assumption been systematically tested and confirmed? For cultivar rankings? For optimal N rate? For the effectiveness and reliability of IPM regimes? 

Assumption 1. Zone of Inference....that research results derived from research station plots or a plots located on a few farms will be generalizable over some recommendation domain (e.g. southern Ontario; or a given heat unit zone) (Clark et al., 1996).
Assumption 1 diagram

Query. What fraction of all the variability in applied trials is accounted for by main effects vs. interactions, reps, and unexplained variability? If main effects account for a minor share, then just how useful can they be? Does this partitioning vary systematically with level of purchased inputs? Do GxE or MgtxE account for a useful and interpretable fraction of variability, or are interactions so complex and site/year-specific as to defy simple analysis? If the latter, what inferences should be taken regarding the design and conduct of SA research?
Assumption 2. Main Effects.....that parameters of interest respond primarily to main effects. The design of most agronomic and breeding research clearly assumes that interactions with on-farm edaphic, topographic, use history, and managerial factors are of secondary or minor importance in accounting for variability in economically important responses. Or from another perspective, that the genetics of modern cultivars/hybrids are assumed to be sufficiently robust and widely adapted as to be little affected by environment or management.

These assumptions may be valid for farmers willing to purchase the inputs intended to reduce constraints to production, homogenize growing conditions, and hence, mimic research station conditions. But as purchased inputs are replaced with system design decisions on SA farms, natural levels of environmental heterogeneity may be more fully expressed. A less homogeneous cropping environment, in time and space, may imply a greater scope for yield-determining interactions among inputs, management, and environment on SA farms. As such, research to support SA may be characterized by a greater emphasis on interactions among main effects and the multiplicity of variables likely to pertain in on-farm practice.

Assumption 3. Independence .....that enterprises are best studied in isolation from other enterprises (Table 1). True enough, from the perspective of publishability. But DO enterprises actually behave independently on commercial farms? When assessing response to N in wheat or to rootworm insecticides in corn, the influence of the preceding crop or crop series is commonly acknowledged. But what about everything else?

Query. Does the "information" imbedded in a given field or plot , including but not limited to the preceding and neighboring crops, affect the responsiveness of the current crop to ....whatever? And hence, the generalizability of findings from one site to another? 

For example, how does the "type" of preceding crop affect the seasonality of N mineralization in subsequent years? How does the aggressiveness of a winter cereal in spring affect the weed seed bank next spring?

The essence of pest-control in SA, as against input-based agriculture, may very well be the particular choice and sequencing of crops and crop operations (cultivation, compost application, cover cropping) in the years leading up to a given crop/year. Perennial weed communities, the weed seed bank, and the soil/residue reservoir of both pests and pathogens as well as deleterious soil microbes are all known to be influenced by
Table 1. Assessment of the degree to which NE SARE projects (1998 Progress Report) have addressed interactions within vs. among enterprises (excluding 37 others dealing primarily with processing, marketing, CSA's, compost/soil health, conferences/training/videos/manuals, and medicinal herbs
Fruit or Vegetable Crop Grain or Herbage Crop Crop/Livestock Interactions (n=2) System or Watershed (n=9) 
Individual Crops (not interacting) (n=24) Interactions Among Two or More (n=0) Individual Crops (not interacting) (n=5) Interactions Among Two or More (n=2)
p.17; Murray p.23; Cherney p.13; Cox p.37; Hartsock p.33; Wonnacott
p.21; Murray p.28; Cox p.27; Jokela p.161; Wolfe p.35; Griffin
p.55; Berkett p.29; Heckman p.39; Duesterberg
p.61; DeMoranville p.30 Salon p.41; McCrory
p.65; Averill p.59; Drummond p.109; Zweigbaum
p.67; Prokopy p.113; Holden
p.69; Fiola p.123; Dengler
p.71 Drummond p.153; Lamont
p.73; Travis p.155; Brumfield
p.75 Shearer
p.77; Kovach
p.79; MacHardy
p.81; Cooley
p.82; Polavarapu
p.147; Hoffmann
p.151; Hazzard
p.157; Bellinder
p.159; Bjorkman
p.63; Hackman
p.165; Cogwill
p.167; Petzoldt
p.169; Heckman
p.172; Ashley
p.173; Orzolek

the history of crop type and associated management practices(1) (Bezdicek and Granatstein, 1989; Dick, 1993; Reganold, 1995). Does it follow, then, that pest and pathogen control or nutrient dynamics in a given crop can be usefully studied in a vacuum created experimentally to control exogenous variation and enable detection of main effect trends?

Within the projects undertaken in the northeastern SARE program (reported in the 1998 Progress Report), interactions among crop and livestock enterprises (n=2), among grain/herbage crops (n=2) and among fruit/vegetable crops (n=0) are particularly scarce (Table 1). Many solid and laudable efforts are being made to integrate within a given crop, focusing on crop-pest-pathogen-management interactions. However, a closer look at interactions among enterprises, and particularly between crop and livestock enterprises (Table 2), might reveal promising and novel insights into both economic and ecological sustainability.

Assumption 4. Agriculture is inherently unnatural, and problems happen... and our job in research is to solve the problems. Consider an alternative perspective - that our job - and assuredly, that of the farmer - is not to solve problems, but to avoid them in the first place, for both economic and environmental reasons. The dichotomy between solving and avoiding is the difference between farming systems founded upon linear and holistic thinking(2).

Query. To what extent are pest, pathogen, and nutrient problems "management-induced"? Do we really need to add more fuel to that endless cycle of co-evolving resistances by testing yet more insecticides - natural or unnatural - if the real cause of pestiferous populations of CPB or corn rootworm or apple scab is managerial? 

Table 2. Possible benefits of re-integrating crop and livestock enterprises on the same mixed farm, recognizing livestock as tools of production
Feature Enabled by Mixed Farming Whole Farm Benefits
Perennial swards in crop rotation Soil biological diversity and health, organic matter content, tilth, and water infiltration and holding capacity improve; bulk density is reduced; weeds are controlled directly and indirectly; nutrient scavenging can retrieve nutrients which have escaped below the shallow rooting zone 
Livestock as a biological and economic buffer Allows sustainable utilization of land unsuited to annual cropping in space or time. Excess or low-grading grain, cull potatoes, or other residues may be converted into maintenance or growth. Livestock are effectively a "value-added" buffer against vagaries of the weather and marketplace
Nutrient cycling Digestion by livestock accelerates cycling, by putting some elements into more labile form (e.g. N) and by cleaving off digestible C and leaving more stable C chains; livestock can also be used to transport nutrients within the farm, from places of high to low content
Supply of manure Increased soil biological activity, organic matter, and slow release nutrient supply, promotive of healthy crop growth
Crops as nutrient sinks Manure return to the field from which the feed or pasture was grown completes the cycle, reducing nutrient management issues
Cropland and livestock health Stress-induced illness is lessened when stock are outside on pasture for most or all of the year
Sheep pastured in an apple orchard, using MiG techniques Maintain a vigorously regrowing, dense sward that nourishes the sheep, eliminates erosion/degradation on sloping land, and resists weed encroachment. The managed grassland pumps fresh C and N into the soil, sustaining high levels of earthworms while retarding rodent nesting and girdling damage. Sheep consume dropped apples, eliminating next year's codlin moth reservoir. Earthworms consume fallen leaves under the snow during winter, controlling the inoculum source for next year's leaf blight. Each enterprise is complementary to and supportive of the other.

To illustrate the commercial implications of this difference, consider the perspective of Roy Berg and Mick Price, then Dean and Associate Dean of the College of Agriculture and Forestry at the University of Alberta and both respected cattlemen. In an article in the February 1992 "special calving issue" of Cattleman magazine, an issue which is traditionally devoted to producer approaches to calving problems, they said:

"A visitor from outer space reading the February calving issue of Cattlemen for the past few years could be forgiven for thinking that pregnancy is a disease that breaks out seasonally in our cow herds and can be only be cured by an invasive procedure called 'calving'."
Berg and Price argued that many of the problems addressed by producers were in fact caused by simply layering one linear technology upon another without addressing the root cause of the problem in the first place (Table 3). Rather than accept that difficult calving was caused by a mistaken "first generation" decision, producers tend to apply successive generations of technology, each of which is not only costly - in money and time - but which in turn creates yet more problematic outcomes. "Pretty soon, we've got technology piled on top of technology trying to solve problems that we wouldn't have if we had carefully examined the technology we were using in the first place" (Berg and Price, 1992).
Table 3. An illustration of the implications of applying linear thinking to a biological system
Decision Technology Goal Outcome
First Generation selection of a bull that throws large calves, with calving early in the year to increase fall weaning weights difficult calving in unnaturally cold and stressful conditions
Second Generation improved calf puller to extract the calf from the cow weak and listless calf
Third Generation warming box, calf mover, colostrum warmer etc. to increase calf survival bonding problems

Is it so implausible that our incessant search for ways to control pest, pathogen, and weed problems personifies the same costly pattern- a mistaken pattern which must be broken if we are to truly serve the needs of sustainable agriculture?
Many agricultural pests are not born but are created by management decisions that create a niche for them to proliferate to pestiferous proportions. To a very real degree, dependence upon purchased inputs to "solve" problems caused by linear thinking(3) is the reason why contemporary agriculture has such narrow margins. Narrowing margins, of course, are the prime driver of the imperative to get ever larger just to stay even, which in turn, creates a wider suite of rural community problems. 

Holistic systems, which are designed to mimic and channel natural processes, have the potential for greater profitability. It is economically rational to replace natural processes with synthetic inputs only when the cost of the inputs is less than the value of the resultant commodities - a reality that hasn't pertained in years for many crops.

Assumption 5. Science is Objective and Value-Neutral. This is an argument that is increasingly difficult to make with a straight face, particularly after the recent series of revelations in genetic engineering (Table 4). Yet this increasingly tired horse is lamely trotted out in a disturbingly transparent attempt to rebut challenges to the conventional paradigm.

Query. What is the likelihood of finding something if you are not looking for it? How would a research program differ if it were to start from the premise that sweet corn or tomatoes or potatoes would be grown not just with IPM or reduced inputs, but without biocides entirely? 

If science in general and agricultural science in particular were truly objective, then why is there only "one way", with even the suggestion of "alternatives" regarded as a direct affront by so many researchers? Why does the ag research sector go ballistic at reports that soils managed organically, using tillage, can actually be higher in OM and healthier than those using reduced tillage and herbicides (Reganold, 1988 and 1995; Weil et al., 1993)? What is so very threatening - not just to those with a vested economic interest, but to independent, publicly funded university faculty and government researchers - about the finding that milk can be produced with less environmental impact, less mastitis, less family stress, and more producer profit on grass than in confinement?

The claim of scientific "objectivity" may in fact be a mantle to deflect or obscure the possibility of alternative approaches to agricultural production. In the simplest sense, a scientist who is truly "objective", which my aged dictionary defines as "free from or independent of personal feelings, opinions, prejudice; detached; unbiased", and applies good scientific method to a problem can derive only one objective solution- right? By default, then, if someone comes up with another solution, then they must either not be objective, or must not have done very good science - right? Nice, tight, exclusionary reasoning, but it starts from the premise that scientists really are able to elevate themselves and their research completely from the values that drive every other facet of their lives. What is wrong with this picture?

If we want to call ourselves objective, then we had better start earning the adjective. We need to get past the status-quo values implicit in "conventional wisdom" in order to test and "see" alternatives. An anecdote recalled in Joel Salatin's new book You Can Farm (1998) puts it bluntly. Regarding agricultural researchers who challenge the feasibility of small-scale, alternative farming, an observer states "they used to say 'I'll believe it when I see it'. Now they say 'I'll see it when I believe it."

It is tempting to just recoil in disbelief from such a harsh indictment, but on the other hand, why not take a lesson from it? I cannot think why scientists should be more objective than anyone else - we are, after all, only human. But at the least, the values which drive our work, whatever they are, should be acknowledged by us and be compatible with those of our clients.
Table 4. Is science objective? Questions from the field of genetic engineering
Issue References
Absence of scientifically sound, experimentally verifiable resistance management plan for plant pesticides (e.g. Bt) and for herbicide resistance cultivars (e.g. RR) Benbrook and Hansen, 1997 (summarize the evidence that the "high dose-refugia" scheme won't work

Gressel, 1996 (debunks the then prevailing myth that glyphosate resistance is impossible)

Ives, 1996 (provides evidence that refugia won't work)

Johnson and Gould, 1992 (show that Bt transgenes will accelerate resistance to Bt)

Tabashnik et al., 1997 (demonstrate that frequency of Bt resistance alleles is at least 10X higher than assumed, that the resistance alleles are retained in the absence of selection pressure, and that the same allele confers resistance to four different endotoxins)

Lack of consideration of ecological ramifications of GE interventions Donegan et al., 1997 (show that transgenic proteinase inhibitor I - an insecticidal protein - in buried GE tobacco residues alters the species composition of the soil biota responsible for organic matter decomposition and nutrient cycling)

Holmes and Ingham, 1995 (discuss how GE organisms could affect a soil foodweb)

Tapp and Stotzky, 1995 (demonstrate that the active BT-endotoxins synthesized by Bt-plants can persist for extended periods in the soil, bound to clay particles where they are resistant to breakdown by soil microbes). 

Lack of consideration of horizontal gene flow implications Ho and Tappeser, 1996 (document the movement of genes among quite different bacteria, between fungi, between bacteria and protozoa, between bacteria and higher plants and animals, between fungi and plants, or between insects).

Hoffmann et al., 1994 (report the movement of antibiotic resistance genes from GE-rapeseed, black mustard, thorn-apple, and sweet peas into a soil fungus Aspergillus niger)

Stephenson and Warnes, 1996 (conclude that at least in theory, any genetically engineered trait can be transferred to any prokaryotic organism and many eukaryotic ones) 

Assumption that transgenes cause only the intended effects, and do so consistently Doyle et al., 1995 ( report that Pseudomonas engineered to degrade 2,4-D also affect soil fungi)

Myerson, 1997 (reports on the issue of boll drop caused on 25% of the land sown to RR cotton in Mississippi, and subsequent successful lawsuits)

Absence of consideration of genetic pollution MacArthur, 1998a and b (documents movement of transgenes among canola fields in the prairies)

Rifkin, 1998 (outlines various criticisms of GE, including lack of catastrophic insurance by life science companies


  1. While an essential prerequisite to sustainability, profitability is a distinct issue which - for many reasons - is better addressed at the level of policy than intertwined with sustainability itself.

  3. To genuinely support producers seeking to adopt more sustainable farming practices, the principles underlying ecological sustainability need to be enunciated as the foundation upon which all other research rests.

  5. Research to support ecologically sustainable agriculture is likely to differ not just in substance, but in methodology. Some of the needed changes are revealed by critically assessing some of the assumptions which have underlain much of conventional, inputs-based research for the last several decades.


Benbrook, C.M. and M. Hansen. 1997. Return to the "Stone-Age of Pest Management". Address before the EPA Public Meeting "Plant Pesticides Resistance Management", 21 March 97, Washington, D.C.

Bezdicek, D.F. and D. Granatstein. 1989. Crop rotation efficiencies and biological diversity in cropping systems. Amer. J. Altern. Agric. 4:111-116

Clark, E. Ann and S.F. Weise. 1993. Ch. 11. A forage based vision of sustainable agriculture. pp. 95-110, in: T. Simms (ed) Agricultural Research in the Northeastern United States: Critical Review and Future Perspectives. American Society of Agronomy, Madison, WI.

Clark, E. Ann, B.R. Christie, and S.F. Weise. 1996. The structure and function of agricultural research. Can. J. Plant Sci. 76:603-610.

Dick, R.P. 1993. A review: long-term effects of agricultural systems on soil biochemical and microbial parameters. In: Paoletti and Pimentel (ed) Biotic Diversity in Agroecosystems Elsevier Press, Amsterdam. 356 pp.

Donegan, K.K., R.J. Seidler, V.J. Fieland, D.L. Schaller, C.J. Palm, L.M. Ganio, D.M. Cardwell, and Y. Steinberger. 1997. Decomposition of genetically engineered tobacco under field conditions: persistence of proteinase inhibitor I product and effects on soil microbial respiration and protozoa, nematode and microarthropod populations. J. Applied Ecology 34:767-777.

Doyle, J.D., G. Stotzky, G. McClung, and C.W. Hendricks. 1995. Effects of genetically engineered microorganisms on microbial populations and processes in natural habitats. Adv. Appl. Micro. 40.

Faeth, P., R. Repetto, K. Kroll, Qi Dai, and G. Helmers. 1991. Paying the Farm Bill: US Agricultural Policy and the Transition to Sustainable Agriculture. World Resources Institute.

Faeth, P. 1995. Growing Green: Enhancing the Economic and Environmental Performance of U.S. Agriculture. World Resources Institute.

Gressel, J. 1996. Fewer constraints than preoclaimed to the evolution of Glyphosate-resistant weeds. Resistant Pest Management 8(2):2-5.

Ho, Mae-Wan and B. Tappeser. 1996. Transgenic transgression of species integrity and species boundaries. Presented at the Workshop on Transboundary Movement of Living Modified Organisms Resulting from Modern Biotechnology, Aarhus, Denmark, 19-20 July, 1996. http://userwww.sfsu-edu/~rone/GEEssays.html

Hoffmann, T., C. Golz, and O. Schieder. 1994. Foreign DNA sequences are received by a wild-type strain of Aspergillus niger after co-culture with transgenic higher plants. Curr. Genetics 27:70-76.

Holmes, T.M. and E.R. Ingham. 1995. The effects of genetically engineered microorganisms on soil foodwebs. Supp. Bull. Ecol. Soc. America 75/2

Ives, A.R. 1996. Evolution of insect resistant to Bacillus thuringiensis-transformed plants. Science 273:1412-1413.

Johnson, M.T. and F. Gould. 1992. Interaction of genetically engineered host plant resistance and natural enemies of Heliothis virescens (Lepidoptera: Noctuidae) in tobacco. Environ. Entomol. 21(3):586-597.

MacArthur, M. 1998a. Canola crossbreeds create tough weed problem. Western Producer, 15 October 98.

MacArthur, M. 1998b. Resistant canola expected. Western Producer, 15 October 98.

Myerson, Allen R. 1997. Seeds of discontent: cotton growers say strain cuts yields. New York Times 19 Nov 97.

Reganold, J.P. 1988. Comparison of soil properties as influenced by organic and conventional farming systems. Amer. J. Altern. Agric. 3(4):144-155.

Reganold, J.P. 1995. Soil quality and profitability of biodynamic and conventional farming systems: a review. Amer. J. Altern. Agric. 10:36-45.

Rifkin, J. 1998. Apocalypse when? New Scientist. http://www.newscientist.com/nsplus/insight/gmworld/gmfood/rifkin.html

Stephenson, J.R. and A. Warnes. 1996. Release of genetically-modified microorganisms into the environment. J. Chem. Tech. Biotech. 65:5-16.

Tabashnik, B.E., Y-B Liu, N. Finson, L. Masoson, and D.G. Heckel. 1997. One gene in diamondback moth confers resistance to four Bacillus thuringiensis toxins. Proc. National Acad. Sciences 94:1640-1644.

Tapp, H. and G. Stotzky. 1995. Dot blot enzyme-linked immunosorbent assay for monitoring the fate of insecticidal toxins from Bacillus thuringeiensis in soil. Applied and Environmental Microbiology 61(2):602-609.

Weil, R.R., K.A. Lowell, and H.M. Shade. 1993. Effects of intensity of agronomic practices on a soil ecosystem. Amer. J. Altern. Agric. 8(1):5-14.


1. This cumulative carryover effect may well be the basis of the 3-5 year "transition interval" often observed upon adopting organic methods

2. Linear thinking is "one cause-one effect", while holism is the awareness that everything is connected, such that changes to one element can ramify out and manifest themselves in other facets of the system. To a linear thinker, a patch of weeds is a problem and cultivation or herbicides are the solution. To a holist, the weeds are a symptom of a larger, system problem and the question is "what have I done that created the niche that allowed the weeds to proliferate, and how can I change the niche to disfavor the weeds"?

3. eg. simple pest-friendly rotations driven by market opportunities and government incentive programs rather than ecological sense