INTRODUCTION

The need to improve tree performance and productivity for diverse products is widely acknowledged. Agenda 21 of UNCED (Chapter 21 and the "Forest Principles") re-affirms the need "to promote the identification of more productive strains of fast-growing trees, especially fuelwood species, and the development of rapid propagation methods to aid their wider dissemination and use".

A similar call to increase crop productivity was made in the agricultural sector in the sixties, following a severe crisis in food production in many parts of the world. Major advances in agricultural research during the 10-15 years that followed, later known as the "Green Revolution", resulted in a massive overall increase in productivity, in which the genetic improvement of crops was a very important component. While the increase in productivity was impressive, the "Green Revolution" has been criticised for the inability of established mechanisms to facilitate adoption of the new varieties in remote regions. It was also claimed that only rich and/or large farmers were able to use and profit from the new varieties, because they generally required high-input cultivation methods. Ineffective seed production and distribution systems were also considered to have limited the spread of improved varieties at the farm level (Dalrymple and Srivastava, 1994).

The use and role of trees on farms is gaining more attention as systems of agroforestry and farm forestry are strengthened. In general, trees integrated into farming systems receive less attention and management inputs than agricultural crops, but in some areas farmers are now managing trees more intensively than before, bearing the associated costs in labour and materials.

To ensure that the economic and environmental benefits which can accrue from the use of higher quality and better adapted tree germplasm are widely realised, we must examine the processes by which such germplasm is distributed to and subsequently moved among smallholder farmers, as well as other tree growers. To achieve this we believe it instructive to study the experiences in relation to diffusion (the term "diffusion" is used in this paper to signify the process of seed movement in general, rather than the use of a formal distribution system) of improved germplasm of agricultural crops. We also believe that additional efforts should be made to study patterns of movement and exchange of germplasm of tree species preferred by smallholder farmers.

STEPS IN THE PROCESS OF DIFFUSION

The successful diffusion of better adapted and genetically improved germplasm relates mainly to three problem areas. These are: (a) the availability of such germplasm, referring to its identification, development and production, (b) the access to it, and, (c) its rate of adoption and use.

(a) Availability of better-adapted and genetically improved planting material

Improved varieties of a number of tropical agricultural crops have been developed during the past 30 years. A precondition for a wide adoption of improved varieties is that the material is readily available, i.e. multiplied in adequate amounts at the right time of the year for dispersal. The multiplication of germplasm of agricultural crops is often mentioned as a bottle neck for the wide distribution of improved varieties due to inadequate production methods, lack of resources, and poor planning (Heisey, 1990; Morris, 1995).

Multiplication of germplasm of trees differs from multiplication of agricultural crops. For example, trees are perennials and can continue to provide a regular and cheap source of germplasm for a long time from the living tree. The development of methods to multiply tree germplasm (propagation techniques, establishment of seed orchards and seed production areas) is usually integrated into the overall domestication and improvement strategies. One serious restraint, however, is the quantity of seed produced which, for some species, can be much less than the overall demand.

Tree improvement and research in propagation methods is being carried out by national institutions in many tropical countries. In addition some tree improvement networks have been established on a regional basis for example in Asia the FORTIP project (UNDP/FAO Regional Project on Improved Productivity of Man-Made Forests through Application of Technological Advances in Tree Breeding and Propagation (RAS/91/004)) and the ASEAN Forest Tree Seed Centre. It is to be expected that superior germplasm of several tree species will be identified and developed in the near future and the possibilities for enhancing regional diffusion of this material needs to be assessed. Because trees have a much longer rotation period than agricultural crops, it is important to disseminate, as quickly as possible, high-quality germplasm of useful species or, where applicable, improved strains: otherwise available suitable sites may be planted with stock of inferior genetic quality which may occupy the land for many years. To do this we need to know what constraints will affect the diffusion process.

(b) Access to better-adapted and genetically improved planting material

As improved varieties of agricultural crops were developed, formal seed supply systems were set up in many developing countries, often with support from development assistance agencies. Nevertheless, success with the widespread dissemination of improved seed to small farmers through the formal sector seed organisations has been mixed. Limited institutional capacity, lack of funds and resources and inadequate planning, have combined to limit the movement of superior seed to smallholder farmers through the formal seed supply systems. Few developing countries have well-functioning formal seed supply systems, and informal seed diffusion mechanisms are responsible for the majority of seed movement. These informal seed diffusion mechanisms have been under-researched and their potential contribution is frequently overlooked in national seed programmes (Cromwell, 1990).

At a local level, informal seed supply systems can be very complex and are affected by problems such as storage, unavailability of seed, lack of finance, high variation in yield due to climate, pests and other risk factors. It is very common for farmers to use seed retained from the previous harvest. According to a study carried out in 1993, poor farmers in Rwanda bought 90 % of their seed, because they were frequently obliged to use the seed to feed their families (Sperling and Loevinsohn, 1993). A very active farmer-to-farmer seed exchange system has been found in the hills of Nepal (Cromwell et al., 1993). In Central America local seed diffusion systems were found to be very heterogeneous and highly dynamic. New varieties were obtained from local sources and the farmers also used the local seed exchange mechanism for renewal of the seed. Farmers gave their 'tired' seed to farmers in cooler and more fertile areas for rejuvenation. This also served as an off-farm seed storage system. In some studies farmers were also found to select their seed carefully, e.g., by harvesting seed from the centre of the field to maintain purity, or by selecting heads for seed from a range of plant types (Almekinders et al. 1994).

Informal seed supply systems for agricultural crops are complex, and can differ considerably between regions. The questions we need to ask, before setting up strategies for the diffusion process of forest tree seed, are how strong are the existing informal supply systems of germplasm of trees; what complexities are involved; and how and by how much do they differ between regions?

Seeds of most forest tree species are not eaten, and if the seed is consumed/wasted (e.g., fruit trees) vegetative propagation might be used instead. Also the need for rejuvenation of the material might be minimal because seed is renewed and available each cropping season on the tree. On the other hand farmers obtain planting stock through various methods, not only through raising trees using seed collected by themselves. In some cases the majority of trees planted on farms may originate from other sources, such as naturally regenerated seedlings (wildlings) transplanted from elsewhere on-farm or off-farm or from vegetatively propagated cuttings. This includes many fruit trees, as well as multipurpose species which are growing in living fences and hedgerows (Shanks and Carter 1994). Production of seedlings can be an additional source of income for a farmer or the seedlings can be produced in small-scale private nurseries. Existing inferior genotypes or land races of exotic species can be the source of unimproved material for a long time and compete strongly with better-adapted and improved germplasm. This might provide, and probably does, a dis-incentive for farmers to adopt improved and usually more expensive material. In addition the planting material has little intrinsic value unlike seed of most agricultural crops which retain a value for consumption.

We have been unable to find much literature on the mechanisms of supply systems of germplasm of trees at the community level. This lack of information indicates that further research is warranted into the informal supply systems of tree planting materials.

(c) Adoption and use of better-adapted and improved planting material

Because of the limited capacity of the formal seed sector in most developing countries it is often only a small proportion of farmers who have access to new varieties of plants. The rate of adoption has also been found to be constrained by lack of farmers' participation in different extension activities. The level of contact with extension workers is low among smallholder farmers and especially older and/or illiterate farmers. This can be associated with the low level of farm resource endowments, inadequate and untimely supply of farm inputs and lack of economic motivation in terms of higher crop output and its price (Hugar et al., 1992). A study in Pakistan indicated that a low level of contact between farmers and the formal seed system was related also to other variables than farm size such as literacy and extension contact (Tetlay et al., 1991). Cromwell (1990) found that farmers who initially tested new varieties were those with relatively more land and more formal education, and often holding a traditional leadership role in the community. Among smallholders adoption is not a one time or final act. Their use of even well-appreciated varieties may involve repeated cycles of loss and restocking; and adoption may thus require continuing support (Sperling and Loevinsohn, 1993). These patterns of farmer contact with formal seed supply systems and extension would probably also apply to some extent in relation to tree planting and the use of improved tree germplasm.

Even among farmers who have access to improved germplasm and extension services, low levels of adoption of new varieties is a frequent problem. A study in Ghana found the main reasons to be storage problems, marketing problems, cooking quality and yield (Marfo and Tripp, 1991). Only a few are willing to risk growing a new variety before it is widely proven locally (Cromwell, 1990). Farmers need to be convinced that new varieties are truly and consistently superior under their conditions (Borlaug, 1977). Farmers are reluctant to change if the new variety is only marginally better (Bal and Douglas, 1992).

Well-adapted, but less productive, traditional varieties are difficult to displace especially in stressed environments (Dalrymple and Srivastava, 1994).

Financial problems also contribute to the low levels of adoption. Only a few farmers can pay cash for seed. Kelly (1989) found that in many developing countries it will be impossible for the poorer farmer to buy seed before planting time; they require some credit arrangement which allows them to purchase the seed at the right time but to pay for it after the crop has been harvested. The same might also apply to purchase of more-expensive seedlings which have been grown from more costly seed of better-adapted material.

In earlier times the effort to develop improved crop varieties focused on the productivity of the single crop in question not taking into account the interactions of the particular crop in an integrated farming system. New varieties may require earlier planting or earlier harvesting of a previous season's crop on the same land. They may compete for scarce labour at critical points in the production cycles of other crops. A later-maturing variety may introduce a constraint in the family consumption calendar if its longer period in the field coincides with a period when food substitutes are unavailable. It may introduce family cash constraints if delayed harvest prolongs a period of cash shortage. A later-maturing variety may also preclude a second crop the same year. In this situation farmers may choose to plant the combination of varieties that gives the best yield in two growing seasons, as single season yields may be less important than annual productivity (Haugerud and Collinson, 1990).

Although the above examples are related to annual agricultural crops the uncertainties that have been raised might also have some relevance to choice of trees for planting on farms. Are the right species being developed? Have farmers needs actually been considered in the definition of what constitutes an "improved variety"? Does the global concern of the need for increased productivity also consider that people with individual and basic needs are involved; that there are reasons for non-adoption which may not be directly related to the performance of the improved germplasm but to the general management of the farm? The experience gained in the spread of new agricultural crop varieties does emphasise that single-crop or commodity research programmes must not ignore consideration of other crops that compete for farmers' land, labour and cash resources and that help farmers meet their food and cash needs (Haugerud and Collinson, 1990). With reference to the agricultural seed industry, Pray and Ramaswami (1993) express a similar concern: 'policy makers must decide whether the seed industry is an important constraint or whether the so-called improved varieties are actually inferior to the local farmers "varieties".

It seems opportune also to review the relevance of tree improvement research, and to attempt to incorporate the tree planters (e.g. farmers) needs in the efforts to identify species and provenances, and to develop improved germplasm. For example, experiments are often clean-weeded on research stations, but many farmers will lack the time or labour to weed as intensively, and if such interventions are necessary to gain the full potential of the new varieties, then the improved germplasm may not serve the needs of the farmer. Most farmers on small farms do not want trees that compete strongly with crops; slower-growing less competitive species or genotypes are preferred in agroforestry systems. E.g. research in Bangladesh showed that most farmers wanted more fruit and not fuelwood trees although the majority reported scarcity of fuelwood. Farmers in India often wanted a cash incentive such as sale items and would not plant fuelwood for themselves until the already depleted supply was totally exhausted. Furthermore in some areas the key issue was the survival of the trees rather than yield potential (Simons, 1992).

Literature does not offer much help in answering the questions raised in relation to the adoption of the improved or better-adapted tree germplasm. Like the mechanisms of the informal seed supply systems, information on the social structures which determines tree planters behavior and the basis for decisions related to the adoption of new germplasm will have to be collected at local level as differences are likely to exist between regions.

CONCLUSION

There are still several areas to be considered in relation to diffusion of well-adapted and improved germplasm of trees and perhaps also more lessons to be learnt from agriculture that are applicable to the diffusion of germplasm of tree species to smallholder farmers.

Farmers normally work in a time frame that allows them to change in small degrees every cropping season, and to have the possibility to adjust or change in due time. For example, if a crop yield is declining over time and in one year is getting below acceptable and expected level, the farmer knows it is time to exchange or rejuvenate the seed. A tree crop that eventually turns out to be disappointing or even a failure in growth or quality will affect returns for a longer period. Therefore, it represents a much higher risk which farmers with little experience in managing trees have no experience with. To be able to adopt the use of improved and better-adapted, and often more expensive, germplasm of trees, the farmer might have to rely more on the advice and assistance that he can get from extension workers or the formal seed system. For most smallholder farmers this new situation involving uncertainties and dependency of third party is not attractive. Therefore the benefits of the new germplasm will have to be proven and be made highly visible in order to motivate farmers to grow it.

A field of research which has not been discussed in this paper is the issue of extension in regard to how tree planters should be introduced to new germplasm. As the "Green Revolution" progressed, it was soon recognised that adoption of improved varieties alone could realise only a small proportion of the total yield gain that new varieties potentially offered in combination with other inputs or practices. This resulted in the development of the "technology package" approach, introducing both new varieties and improved crop management. This approach was not very flexible, and frequently failed to adequately address smallholder farmers. Farmer participatory approaches became popular during the 1980s. The term, "farming systems research" was introduced covering different methods with the common element that they took into consideration the complexity of the existing, integrated farming systems.

A broader base of information on the tree planters' needs and the underlying reasons for the adoption and use of new, well-adapted germplasm of trees, will enable research in tree improvement to be focused on real problems. Perhaps even more importantly existing seed supply mechanisms and related issues, can be more adequately taken into consideration in extension work based on such information.

REFERENCES

Agrawal, P.K. & Worede W/Mariam (1995). Seed Supply system in Ethiopia. Plant Varieties and Seeds, Vol. 8, 1-7.

Almekinders, C.J.M., N.P. Louwaars & G.H. de Bruijn (1994). Local Seed Systems andTheirImportance for an Improved Seed supply in Developing Countries. Euphytica, Vol 78, 207-216.

Bal, S.S. & J.E. Douglas (1992). Designing Successful Farmer-managed Seed Systems. Winrock International Institute for Agricultural Development. Development Studies Paper Series. January 1992.

Borlaug, N.E. (1977). The Green Revolution: Can We Make It Meet Expectations? In: Proceedings of the American Phytopathological Society, Houston, Texas, 11-12 August 1975, Vol.3, 6-21.

Brown, L.R. & Eckholm, E.P. (1974). The green revolution:Opportunity Lost? In:By Bread Alone,133-146.

Cromwell, E. (1990). Seed Diffusion Mechanisms in Small Farmer Communities: Lessons from Asia, Africa and Latin America. Overseas Development Institute, London, United Kingdom. Network Paper 21.

Cromwell, E. (1993). Economic Reform and the Performance of the Seed Sector in Eastern and Southern Africa. In Technology and Developing Country Agriculture: The Impact of Economic Reform, 113-125.

Cromwell, E. & S. Wiggins with S. Wentzel (1993). Sowing Beyond the State. NGO and Seed Supply in Developing Countries. Overseas Development Institute, London, United Kingdom (134 pp).

Dalrymple, D.G. & J.P. Srivastava (1994). Transfer of Plant Cultivars: Seeds, Sectors and Society. In: Agricultural Technology: Policy Issues for the International Community, 180-207.

Haugerud, A. & M.P. Collinson (1990). Plants, Genes and People: Improving the Relevance of Plant Breeding in Africa. Experimental Agriculture, Vol. 23, 341-362.

Heisey, P.W. (1990). Accelerating the Transfer of Wheat Breeding Gains to Farmers: A Study of the Dynamics of Varietal Replacement in Pakistan. CIMMYT Research Report No. 1.

Hildebrand, P.E. (1988). Technology Diffusion in Farming Systems Research and Extension. Hortscience, Vol. 23 (3), 488-490.

Holding, C., T. Niemi, W. Omondi, S.M. Weru & B.M. Kamondo (1995). Establishment of Seed Production Units in Nakuru and Nyandarua Districts. Technical Report No. 11. Widagri Consultants Ltd. FINNIDA in cooperation with The Ministry of Environment and Natural Resources, Forest Department, Kenya (33 pp + annex 70 pp).

Holding, C. & T. Niemi (1995). Seed Procurement and Distribution. Experiences and Recommendations from Nakuru and Nyandarua Districts. Technical Report No. 16. Widagri Consultants Ltd. FINNIDA in cooperation with The Ministry of Environment and Natural Resources, Forest Department, Kenya (25 pp + annexes 12 pp).

Hugar, L.B., K.B. Umesh & S. Suryaprakash (1992). Constraints in Diffusion of New Farm Technology - An Economic Evaluation. Agricultural Situation in India, Vol 47 (1), 15-20.

Kelly, A.F. (1989). Seed Planning and Policy for Agricultural Production. The Roles of Government and Private Enterprise in Supply and Distribution. Belhaven Press (Pinter Publishers), London, United Kingdom (179 pp).

Marfo, K. & R. Tripp (1991). A Study of Maize Technology Diffusion in Ghana: Some Preliminary Results. In: Africa's Agricultural Development in the 1990s: Can it Be Sustained. Proceeding of a Workshop 14-18 May 1991, Arusha, Tanzania (CASIN/SAA/Global 2000), 56-69.

Maurya, D.M., A. Bottrall & J. Farrington (1988). Improved Livelihoods, Genetic Diversity and Farmer Participation: A Strategy for Rice Breeding in Rainfed Areas of India. Experimental Agriculture, Vol. 24, 311-321.

Moberg, S. (1994). The Research and Seed Project in Zambia. IDRCurrents, 1994 (7), 22-24.

Morris, M. (1995). CIMMYT forges new links with private seed sector. Asian Seed, Vol 2 (6), 12-13.

Palmberg, C. (1986). Selection and Genetic Improvement of Indigenous and Exotic Multipurpose Tree species for Dry Zones. Agroforestry Systems, Vol 4, 121-127.

Pray, C.E. & B. Ramaswami (1993). A framework for Seed Policy analysis in Developing Countries. In: Agricultural and Food Marketing in Developing Countries: - selected readings.

Saxena, N.C. (1994). India's Eucalyptus Graze - The God that Failed. Sage Publications India Pvt Ltd, New Delhi, India (259 pp)

Shanks, E. & J. Carter (1994). The Organisation of Small-Scale Tree Nurseries. Studies from Asia, Africa and Latin America. Rural Development Forestry Study Guide 1. Overseas Development Institute, London, United Kingdom (144 pp).

Simons, A.J. (1992). Genetic Improvement of Non-Industrial Trees. Agroforestry Systems, Vol 18, 197-212.

Sperling, L. Rethinking the Farmer's Role in Plant Breeding: Local Bean Experts and On-Station Selection in Rwanda. International Centre for Tropical Agriculture, Butare, Rwanda. Unpublished (19 pp).

Sperling, L. & M. Loevinsohn (1993). The Dynamics of Adoption: Distribution and Mortality of Bean Varieties among Small Farmers in Rwanda. Agricultural Systems, Vol. 41, 441-453.

Tetlay, K.A., P.W. Heisey, Z. Ahmed & A. Munir (1991). Farmer's sources of Wheat Seed and Wheat Seed Management in Three Irrigated Regions of Pakistan. Seed Science & Technology, Vol. 19, 123-138.