SECTION 1
INTRODUCTION

Chapter 2
ICRAF and a decade of agroforestry development

Howard A. Steppler
Chairman, ICRAF'S  Board of Trustees

In the 1960s and early 1970s there was increasing concern for the forested lands of the tropics (Eckholm, 1976). It was clearly recognized that they were under severe pressure. Some thought that commercial exploitation was the problem; others that fuelwood needs were the culprit; while still others believed that shifting cultivation was the root cause. The president of the International Development Research Centre (IDRQ, located in Ottawa, Canada, engaged Mr John Bene in 1975 to study the problem. Bene assembled a small team in Canada, an advisory committee and recruited experts in the various continents, to prepare studies pertinent to their area. The culmination of these various activities, including extensive travel by Bene, was the publication in 1977 of a report entitled Trees, Food and People (Bene et at., 1977).

Bene and his co-authors recognized that the solution to the problems besetting tropical forests arose from population pressure exerted through the need to produce food and fuelwood. They were prophetic in their choice of sub-title for the report, "Land management in the tropics", for that was precisely the nature of their recommendation, although it was not immediately apparent. In the report, they identified some 23 tropical forestry problems. Of these, nine could be considered as dealing with the more traditional forestry problems. One clearly recognized the need to accommodate agriculture and the remainder encompassed problems related to land use, policy and environmental impact. Bene and his co-authors recognized that the key issue lay at the interface of forestry and agriculture. It is not evident whether they coined the word agroforestry to identify that interface. What is clear, however, is the prominence and widespread use accorded the term since their publication.¹ Their most significant recommendation was to establish an International Council for Research in Agroforestry (ICRAF).

Thus an old practice was institutionalized for the first time.

In the first years of its operation, ICRAF directed its attention to assembling the contemporary knowledge of agroforestry. Several international conferences and workshops were held (Nair, 1987a), of which four are particularly worth mentioning here: one dealing with soils research in agroforestry (Mongi and Huxley, 1979); the second with international co-operation (Chandler and Spurgeon, 1979); the third treated plant research and agroforestry (Huxley, 1983); while the fourth addressed the problem of education in agroforestry. A fifth was held much later and was concerned with land tenure problems.

The Board of Trustees realized by 1980 that, while the collation of information on agroforestry was an important activity for ICRAF, it was not sufficient. ICRAF would need to develop a much sharper focus than envisaged in its charter and mandate if it was to meet expectations. Thus, in 1981, the Board adopted a strategy (Steppler, 1981; Steppler and Raintree, 1983) which set the Council on a path to develop a diagnostic methodology to determine relevance of agroforestry interventions in particular situations. Further, the diagnostic methodology was expected to identify the kind of intervention most appropriate for the situation at hand.

This strategy and focus have served the Council since its adoption. It was based on the Cycle of Technology Development (Steppler, 1981; Steppler and-Raintree, 1983) and is basically concerned with phases I and II of that cycle (Figure 1). In 1984, an external review panel examined ICRAF's total operations. The panel confirmed the wisdom of the choice of strategy and focus when it stated: "The Panel believes that this restricted interpretation (of the mandate) has been appropriate and necessary during these initial years" (Cummings et al, 1984).

Figure 1 The cycle of technology development

The panel went on to recommend that the Council should move into a mode of extending and testing its methodology and assisting in the generation of new technology— essentially phases III and IV of Figure 1.

It should be pointed out that by the time of the external review, it had become clear to the Council, both the staff and the Board, that ICRAF's role was much more than that envisaged by Bene. The research that the Council had undertaken in developing its diagnostic methodology had shown that agroforestry as a land-use system was capable of many beneficial effects and with multi-product output; Bene had been right in his choice of sub-title.

The Council had also initiated activity through its Collaborative Programmes to reach out and to respond to the many requests that it was receiving, both from countries and from donor agencies (Torres, 1987). This was not, however, easy.

As previously mentioned, ICRAF, when established, was the first institution dedicated to agroforestry. Similar institutions did not exist at the national level. It is to the credit of the foresters that agroforestry was seen by them to be an essential development — albeit to secure the forest. The agriculturist did not recognize the situation, since loss of forest was not creating a problem for them; rather, the result was more land for agriculture. Thus, one of the first and critical functions that the Council undertakes when entering an area for a collaborative programme is to nurture at the national level the awareness of the contribution of agroforestry and the need to develop a national mechanism to be a focal point for agroforestry activities. The diagnosis and design methodology has a key role to play in this process.

Agroforestry was beset with much anecdotal material. It was clear that we were dealing with an old practice, but what was not clear, however, was the degree of diversity that might be in use. Thus, one of the projects launched in the early 1980s was systematically to inventory agroforestry systems (Nair, 1987b). The project was announced in Agroforestry Systems and several other international journals. Subsequent issues of Agroforestry Systems have carried articles describing specific systems. Nair (1985) published a first approximation of a classification of described agroforestry systems.

Four facts emerge from this preliminary compilation of systems: first, there is a bewildering array of agroforestry systems worldwide, and we have but scratched the surface; secondly, there are relatively few rigorous experimental data pertaining to performance of agroforestry systems; thirdly, the number of tree species — multipurpose trees — being used in various systems is in excess of 2,000; and fourthly, the systems vary from relatively simple with two or three components, to the complex homegardens which may contain upwards of 50 species plus animals and fish (Fernandes and Nair, 1986; Soemarwoto, this volume).

There is no question but that ICRAF should continue to catalogue agroforestry systems. Only through such activity can we build our body of knowledge relating to current practice. The objective in continuing the inventory is not so much to identify systems with which we wish to experiment. Rather, it is to record the kinds of systems used and where, the rationale for their use, the kinds of output in order that new or modified systems shall be designed to achieve the same goals. This is not to rule out the possibility of introducing systems which would have additional features, for example, halt soil erosion or improve soil fertility.

The problem of lack of experimental data or even of production data for the various systems is a serious gap in our knowledge. This is partly a reflection of the newness of agroforestry — it was until recently literally considered as a subsistence form of land management—and partly that there are essentially no experimental techniques applicable to agroforestry systems. Most statistical techniques and experimental designs have been developed for monocultures. There is even relatively little experimental work done with annual crop mixtures. The closest approximation to agroforestry systems would be found with perennials in forage mixtures, but even this is not as complex as agroforestry where one is dealing with at least one tree species and a crop species. ICRAF recognized this problem and has for some years been investigating different experimental designs at its Field Station. Further, it has published several working papers,² many of which deal specifically with this problem.

An additional dimension is the question of appropriate impact/ output measurement. Monoculture and forage mixtures are relatively easy to measure — the output is clearly identified. With an agroforestry system we have multiple outputs. That of the agricultural component will probably be easy to measure—yield being the most visible output. The tree component may be much more elusive. There will be visible outputs, for example, fuelwood, building poles, fruits; others which affect crop production, such as leaf mulch and fixed nitrogen, and yet other nearly invisible ones which could have an effect on the entire system, such as recycling of nutrients from subsoil, control of erosion, increased infiltration rates.

The problem has several dimensions: first there is the need to determine which outputs/ impacts shall be measured; secondly, the need to specify the baseline against which measurements will be made; thirdly, since many outputs and particularly impacts are liable to be qualitative, the need to quantify all measurements; and fourthly, the need to develop some common quantified measure of output which can be applied to any system in order that systems can be compared. For the latter, the most obvious such measure is the economic return. While this should be one of the measures, it is urgent to establish some other measure of the "value" of a system, such as the constancy or sustainability of the system.

The third factor identified was the plethora of multipurpose tree species that are candidates for experimentation. In many respects multipurpose tree research is in a very primitive state compared to agriculture. There is one species of the genus Leucaena (see Brewbaker, this volume) which is relatively well studied, although it pales when compared to wheat or maize — and this is but one out of literally hundreds of candidate species. The great majority of the species are represented by a single collection; there is virtually no information on the genetic variability which exists within a species. Thus, for example, there is a small stand — about 100 trees — of Acacia albida on the ICRAF Field Station at Machakos, Kenya. This stand was established from a seed lot and shows great variability in rate of growth, type of growth, retention of leaves, rate of leafing out—a mere indication of the wealth of variability which probably exists in the species. The same is no doubt true of many others, either within a provenance or between provenances.

There are two major tasks facing us with respect to the multipurpose tree dilemma. The first task is rapidly to screen the hundreds of species and genera and to classify them on a limited number of criteria. 'These criteria must be agreed upon by all researchers undertaking the task. This screening should also have an objective to identify the most promising candidates within each class/group for more intensive study. The second task is to initiate the more detailed studies of the selected candidate species arising from the first task. Again, we must be pragmatic in our approach since we cannot indulge in the luxury of exhaustive physiological studies or cytogenetic research—that will come as the need arises. Work at this stage would be accelerated if one sought a specific ideotype of multipurpose tree to fit a particular niche. The challenge is great and the need for ingenuity and pragmatism in pursuing the research is most desirable — indeed essential. As an input to this whole process, two recent publications are most helpful.

Poore and Fries (1985), in reviewing the status of Eucalyptus as a candidate for agroforestry, indicate the kinds of questions which must be answered for Eucalyptus and, hence, for any candidate species. In so doing, they also demonstrate the difficulty in finding discrete unequivocal answers. Beer (1987), in discussing shade trees for three commercial tree crops (coffee, cacao and tea), lists 20 disadvantageous, 16 advantageous and 21 desirable characteristics for shade trees. Again, many of these are pertinent for any multipurpose tree in an agroforestry system.

The fourth factor is concerned with the very great number of systems, both large and small, which have been identified by the earlier-mentioned Agroforestry Systems Inventory. Some of the concerns and features relevant to this problem were discussed when discussing the second factor. I would like, however, to address a slightly different aspect.

Any one system undergoing experimentation would include, at a minimum, a tree species and a crop species. Each of these could have variation in genotype and management such as spatial arrangement, maturity type for the crop and harvesting methods (e.g., lopping and coppicing timing for the tree). It quickly becomes apparent that we are dealing with a multifactor design with many combinations. As we add species of trees and/ or crops or introduce animals, the experiment grows in size logarithmically. Thus, a single experiment to elaborate the interrelations within an agroforestry system could be very large. Consider comparing systems and we have another almost quantum leap in size. There is a great challenge to develop experimental designs and test methodology applicable to agroforestry research.

The other dimension to the problem is the fact that we have combined long-lived woody perennials with annuals, short-lived perennials and/or animals. Ideally, experiments should continue for the life of the longest-lived component; this could be upwards of 40 years and we cannot wait that long. Thus, we must also devise tests and methods of prediction which will have acceptable levels of confidence in predicting long-term effects. In my judgement, the volume of agroforestry research will increase with time and in the near future there will be the need for these statistical tools.

To date, most of the agroforestry species are known on the basis of one or very few collections (von Carlowitz, 1986). Evidence from economic crops suggests that higher productivity will be obtained from non-indigenous species rather than indigenous ones (Harlan, 1959).

Thus, two consequences are suggested: first to increase the number of collections evaluated for each major candidate species, sampling over as wide a range of environments as possible. The other conclusion that I would draw is that we should, if at all possible, test species from other continents. There should be an active exchange programme of agroforestry germplasm initiated as soon as feasible. This latter means also that suitable quarantine provision should be in place to facilitate the safe international movement of plant material.

It is evident from this brief analysis of ICRAF over the past decade that there have been three milestones for it and for agroforestry. First and foremost, ICRAF was formed in 1977 — agroforestry was institutionalized. Secondly, ICRAF adopted a rigorously defined strategy in 1981; this set it on a very clear path of development within the discipline of agroforestry. And, thirdly, ICRAF, in 1984, was given a clean bill of health by the external review panel and commended for the work accomplished up to that date. It was urged to extend its activities to the many countries seeking its assistance, and this is what it is pursuing vigorously as it enters its second decade. I have discussed many research issues which have arisen during the past ten years. I should now like to address some other issues which have not yet been mentioned and, finally, to suggest some courses of action.

As has been mentioned, agroforestry was initiated by foresters. For that reason, and in addition because it deals with trees, it is frequently inextricably enmeshed with forest policy. This is particularly true if agroforestry is deemed by our diagnostic and design analysis to be the most appropriate land-use system for an area within a forest mandated area. This may even be true when trees are planted on non-mandated land. Forest policy almost invariably runs counter to an envisaged agroforestry use with cropping and/or livestock. The major context of forest policy is conservation and protection with industrial wood as the product. Hence, continuous outputs as expected from agroforestry are contrary to such policy. There is a great need to study forest policy with a view to suggesting amendments which would meet the objectives for forestry and not hinder agroforestry. ICRAF could play a central catalytic role here.

Policy issues also extend to such matters as marketing of the expected products, with pricing policies which will encourage the use of desirable practices. Land tenure must also favour the use of long-term land-use systems which tend to be sustainable and, hence, encourage the farmer to invest labour in trees. ICRAF undoubtedly has a role to play here in this general debate by objectively assessing various policy and tenure alternatives.

Educational and training programmes in agroforestry are key requirements as we move into this second decade of agroforestry development. There are an increasing number of universities, both in the tropical areas and in the temperate, that are beginning to offer programmes in agroforestry; some are at the undergraduate and some at the post-graduate level. Further, and most encouragingly, students from both agricultural and forestry backgrounds are beginning to prepare for careers in agroforestry. As was stated earlier in this paper, ICRAF held a workshop on education in agroforestry in December 1982, the proceedings of which will soon be published. Sufficient to say that there is no clear picture of the optimum curriculum in agroforestry and such may, in fact, never emerge. Rather, we shall probably have three avenues of development for agroforesters: one would be via an undergraduate programme in agroforestry, which might be in a faculty of forestry or of agriculture; a second could be via an undergraduate degree in agriculture with postgraduate training in agroforestry; and a third could be the converse, namely undergraduate in forestry and post-graduate in agroforestry. Of course, any person with a first degree in agroforestry might also pursue post-graduate studies. There is no basis upon which to judge one avenue as being better than another, although certain situations might favour a particular sequence of study.

ICRAF has as part of its mandate a responsibility to provide training in agroforestry. It has already provided much training in the use of its diagnosis and design methodology (Zulberti, 1987). As it moves into a collaborative research mode, it will expand its training offerings to include research methodologies and experimental techniques. The majority of these training offerings are relatively short courses and not designed to provide advanced degrees.

The discussion has been concerned with the preparation of people for a career in agroforestry research. Of equal or possibly greater importance is the need to prepare people for the transfer/ extension of agroforestry technologies as they become available. The problem of education of these people has not been addressed. They will face complex situations where solutions will involve appropriate MPTs, crops, their combined management and spatial arrangement and possibly even animals. It is expecting a lot to anticipate that one person could be prepared and have at his/ her fingertips the bank of data needed to respond to the challenges. Are we looking at a generalist in the field backed up by a group of specialists responding to his requests, or can we even anticipate a generalist with a computer terminal linked to a data bank which will provide the answers to the complex questions—the data bank being constantly updated by researchers? Is it possible that each extension worker is trained in micro diagnosis and design methodology? These questions require consideration now. Although extension is outside of ICRAF's mandate, some agency should at least initiate the process of examination — and ICRAF would seem most appropriate.

Trees often have deep cultural significance and, hence, their retention and management may not be based entirely on pragmatic decisions. Farmers who live on the knife edge of success or failure at the whim of climate and pests often have different priorities from those of the scientist. It is essential that in dealing with these complex agroforestry systems the socio-economic and cultural factors are recognized. Evaluation of these is a part of the diagnosis and design approach developed by ICRAF (Raintree, 1987). One must ensure, in addition, that these same criteria are considered at various stages of development and testing of new technology.

There is a massive research agenda in agroforestry — not only what has been set down in the preceding paragraphs of this paper, but also to be found in the succeeding chapters of this volume, where the various authors have frequently directly or indirectly indicated problem areas pertinent to their topic. There is much more than can be done by any one institution, and certainly much beyond the capacity and mandate of ICRAF as envisaged over the next decade. To move forward and to meet the expectations means that there must be a sharing of responsibilities for the agenda and a pooling of resources and of information for the sake of expediency and efficacy.

An examination of the research requirements, both within and without the agenda, clearly indicates that the activities can be classified according to whether they are:

1. Appropriate for an international or regional organization that the information resulting from the research transcends national boundaries, tends not to be location-specific; or

2. Appropriate for a national organization since the results are location-specific and immediately transferable to the ultimate user, the farmer.

The national research systems do not at this time have specific agroforestry research units, although as a consequence of the ICRAF activity in collaborative programmes many are developing agroforestry co-ordinating committees. However, at the national level any or all of the sectors of the agricultural research system — crops or animals, the forestry research sector (particularly tree nursery activities) and the university — may take part in the agroforestry research programme. All activities from location-specific to basic non- location-specific can be carried out at the national level. However, every national system should undertake the testing at farm level of new systems and the validation of components of the system proposed for its use. Any additional work "up-stream" from this adaptive/applied research would depend on resources available for its use.

At present there are two groups of institutions on the international scene: ICRAF, which constitutes the first group of those dedicated to agroforestry, and the international agricultural research centres, lARCs, which constitute the second group.

This latter group is composed of some 13 centres, nine of which deal with primary production of food commodities with either a regional or world responsibility. This group of 13 is funded by the Consultative Group for International Agricultural Research (CGIAR) (Baum, 1986). There are many other international and regional organizations (for example, ICRAF) outside the CGIAR but funded by the international community and engaged, or with a potential for engagement, in agroforestry. One such organization is the International Union of Forestry Research Organizations (IUFRO) which has held several sessions on agroforestry at its international conferences.

The discussion of needs in agroforestry has centred on the tree component and on the systemper se—the much neglected and virtually unknown subjects. As we begin to refine our information, we shall be seeking genotypes of the other components, particularly of the crops to better fit the system. These crops are in most cases the commodities which are the mandate of the previously mentioned lARCs. Thus, there will be added demand to the lARCs to develop these appropriate genotypes. Further, there are some specific systems such as alley cropping (see Kang and Wilson, this volume) or the interaction between animals and browse, which would best be done by the appropriate lARCs.

Turning now to the research on multipurpose trees per se, it would appear that ICRAF is the most appropriate institution to co-ordinate the activities if not actually to undertake them. As with any plant species, there is much basic work to be done upon which to build the more applied research. It is the former, along with the development of the relevant methodologies, which seems most appropriate for ICRAF.

There are three other functions which are also critical and which would most naturally fall within the ambit of ICRAF.

The first of these is to act as the focal point for information emanating from the various research activities. This is not new to ICRAF, but its central role in this must be reinforced. As activities multiply at the national and international level, this role will increase in importance. Naturally, the concomitant activity is the dissemination of that information to the users.

A second role is to ensure that new and essential areas/problems in agroforestry research are addressed. This means maintaining a constant watch on developments in both the research field and in the application field to identify these new challenges. One would envisage that the continuous and logical use of the "Cycle of Development" (Figure 1) with a constantly improved diagnosis and design methodology would be a major source of such information. The problems uncovered could conceivably run the gamut from the biological to the socio-economic problems, from policy to sustainability to statistical techniques.

Thus, the third role would be to seek out the partners to undertake the investigations identified in role two. Many of these will be highly specialized areas of investigation, with some of the problems having a global connotation, while others are more regional. It would, therefore, seem expedient to begin the process of anticipating the kinds of problems likely to be encountered and to seek the collaborative partners for the undertakings — one might even initiate some preliminary studies.

Finally, before euphoria completely clouds rational thinking and attainable expectations for agroforestry, let us return to reality. Agroforestry will not save the world—it is not the panacea for the ills of land misuse. There are undoubtedly many benefits to be gained from an agroforestry intervention but there may—in fact probably will—be costs. Labour requirements may be higher, production of some selected component may drop, new problems, such as bird damage, might even emerge. The title of ICRAF's first publication, The Wasted Lands (King and Chandler, 1978) may have both raised expectations that agroforestry would correct the problems of these areas and at the same time denied to ICRAF opportunities to work in other areas. In either case, ICRAF has moved beyond these boundaries, but now let us give it the opportunity in this second decade to prove its capability to address and correct problems, remembering that it cannot be all things to all people.

REFERENCES

Baum, W. 1986. Partners against hunger. Washington, D.C.: International Bank for Reconstruction  Development.

Beer, J. 1987. Advantages, disadvantages and desirable characteristics of shade trees for coffee, cacao and tea. Agroforestry Systems 5: 3-13.

Bene, J.G., H.W. Beall and A. Cote. 1977. Trees, food and people: Land management in the tropics. Ottawa: IDRC.

Chandler, T. and D. Spurgeon (eds.). 1979. International cooperation in agroforestry. Proceedings of an Expert Consultation. Nairobi: ICRAF.

Cummings, R.W., J. Burley, G.T. Castillo and L.A. Navaro. 1984. Report of the External Review Panel of the International Council for Research in Agroforestry, September-December,1984. Nairobi: ICRAF.

Eckholm, E.P. 1976. Losing ground: Environmental stress and world food prospects. New York: Norton and Co.

Fernandes, E.C.M. and P.K.R. Nair. 1986. An evaluation of the structure and function of some tropical homegardens. Agricultural Systems 21: 179-210.

Harlan, J.R. 1959. Plant exploration and the reach for superior germ plasm for grasslands. In H.P.Sprogue (ed.), Grassland Publication 53, American Society for the Advancement of Science, Washington, D.C.

Huxley, P. A. (ed.). 1983. Plant research and agroforestry. Nairobi: ICRAF.

King, K.F.S. and M.T. Chandler. 1978. The wasted lands. Nairobi: ICRAF.

Mongi, H.O. and P.A. Huxley (eds.). 1979. Soils research in agroforestry. Nairobi: ICRAF.

Nair, P.K.R. 1985. Classification of agroforestry systems. Agroforestry Systems 5: 97-128.

——1987a. International seminars, workshops and conferences organized by ICRAF. Agroforestry Systems 5: 375-382.

——1987b. Agroforestry systems inventory. Agroforestry Systems 5: 301-318.

Poore, M.E.D. and C. Fries. 1985. The ecological effects of eucalyptus. FAO Forestry Paper 59. Rome: FAO.

Raintree, J.B. 1987. The state of the art of agroforestry diagnosis and design. Agroforestry Systems 5:219-250.

Steppler, H. A. 1981. A strategy for the International Council for Research in Agroforestry. Nairobi:ICRAF.

Steppler, H.A. and J.B. Raintree. 1983. The ICRAF research strategy in relation to plant science research in agroforestry. In P.A. Huxley (ed.), Plant Research and Agroforestry. Nairobi: ICRAF.

Torres, F. 1987. The ICRAF approach to international co-operation. Agroforestry Systems 5: 395-410.

von Carlowitz, P.O. 1986. Multipurpose tree and shrub seed directory. Nairobi: ICRAF.

Zulberti, E..1987. Agroforestry training and education at ICRAF: Accomplishments and challenges. Agroforestry Systems 5: 353-374.

Footnote__________

¹ It is interesting that the term "agroforestry" does not appear in the titles of the 54 works cited by Bene et a/.; rather, "agrisilviculture" is used.

² As of 1 March 1987, some 48 working papers have been produced by ICRAF. These cover topics such as the diagnostic and design methodology, experimental techniques, economic and social studies, bibliographies and soils and agroforestry.