With half of the world's arable land being arid or sensitive to drought, the development of drought­tolerant varieties is an important research aim of plant breeders. But considerable variation exists in the definition of drought tolerance, the crops concerned and the technical approaches. In this article it is argued that drought­tolerant plants are in the first place social artefacts: It is the social and cultural environment that determines the choices in research.

World crop production is largely limited by abiotic stress such as temperature, salinity, and aluminium toxicity. Drought, however, is the main abiotic stress, causing not only differences between the mean yield and the potential yield but also causing yield variation from year to year (yield instability).
Globally, about 35 per cent of the arable land can be classified as arid or semi­arid. Of the remainder, approximately 25 per cent, consists of drought­sensitive soils, while even on the better soils in non­(semi­)arid regions drought stress occurs regularly for a short period or at moderate levels. Further, it has been predicted that in the coming years rainfall patterns might shift due to an increase of the global temperature caused by the burning of fossil oils and the corresponding increase in atmospheric dioxides (Greenhouse Effect). Consequently, farming communities in the northern hemisphere could become increasingly dependent on drought­tolerant varieties.

Different research approaches
The approach towards the creation of drought­tolerant crops does not exist. In fact, several research approaches coexist. We will present some different research directions, while stating that they do not necessarily exclude each other. On the contrary, they could be complementary. They are accents in drought tolerance research that hints at the existence of competing social ideas behind the drought tolerance research agendas.

A first example that illustrates the diversity of approaches is that some plant breeders tend to define the drought problem on crop level. Others tend to look for adaptation to drought on the production system level. The first group might look in particular at new biotechnological procedures for developing drought­resistant plants. The second group will emphasize the need to reconsider the practices of farmers in arid regions as a starting point for research programmes.
A second differentiating aspect is that the research might be orientated towards the most important food crops of the international food chain, such as wheat, maize and rice, or to the orphan crops.
A third difference in focus is to direct research to interchange specific characteristics between major trade crops and drought­tolerant crops. In a first approach, useful genetic information from plants fully adapted to arid conditions (such as cacti) is identified, with the aim of transfering this information to current major trade crops (such as wheat). The focus of Pioneer Hi­bred International, USA, on drought­tolerant maize might be considered an example of this approach. In a second approach, plant breeding activities are concentrated on agricultural established species with known drought tolerance qualities (e.g. bulrush millet and sorghum), aiming at their adaptation to the established needs of commercial food production. The activities of the International Crops Research Institute for the Semi­Arid Tropics (ICRISAT), India, on sorghum, pearl and finger millet, pigeonpea, and chickpea are examples of this approach.
Besides, within each approach again specific choices are being made. For example, within the approach focusing on drought tolerance at crop level, plant breeders have to decide which of the adaptive plant mechanisms to drought might be reinforced (see the article by Visser).

Social shaping of researcht
The purpose of this article is to draw attention to the fact that breeders and biotechnologists exercise choice. No drought­tolerant plant will have any significance without a farmer using it. Understanding the use of new agricultural crops implies understanding the social relations of production in a community. The designer of drought­tolerant plants makes assumptions about the social relations of production of the communities in which these plants are developed and used.
Not only social scientists, but also plant breeders and biotechnologists will have difficulty in answering the question on which assumptions the genetic improvements of plants are based. The importance of such a question is clear: The history of the Green Revolution, and of the transfer of agricultural technologies in general has shown that neglecting the social and cultural base of agricultural research can lead to unpredicted social problems, such as an increasing social differentiation.
It goes without saying that political­economic aspects influence plant breeders' choices. Elsewhere, we have analysed the 'multinationalization' of food production, privatization of research issues, interchangeability of agricultural products, and the politics of controlling farming practices at a distance through the supply of new inputs. Institutional factors play a role in establishing research agendas also, such as commercial interests of companies, career perspectives of the involved researchers, self­interested concern by research institutes to command more government funding etc.. However, in this article we will present some social and cultural ideas which might underlie choices in research on drought stress. Yet we know very little about those values, and how they are created. Therefore, the aim of this article is not to present a complete picture: We only hope to stimulate other researchers to consider the particular mix of social relations and cultural ideas shaping the biotechnological drought tolerance research in their region.

Cultural dominance of specific scientific paradigms
To what extent are plant breeding activities influenced (or dominated) by the prevailing scientific paradigms of their period? Does the dominant reductionism of contemporary biology cause breeders to consider only one specific group of solutions, for example, to search for solutions for drought mainly at crop level? This is of direct importance considering the fact that drought tolerance is only plant specific to a limited extent. Agro­ecosystems modelling approaches may be more important than breeding approaches focused on single crop species.
In describing the research problem, most plant breeders emphasize that drought tolerance is such a complex trait, or even no trait at all. It has been pointed out that it varies from crop to crop, from variety to variety, from place to place. In connection with this point, the relationship between drought and other environmental stresses is considered to be a further complication of the research problems of finding the appropriate drought tolerance gene­structure.
Variations in crop performance seem to be considered as the main bottlenecks in finding general solutions for the adaptation of crops to drought. But within another (more socially­orientated) paradigm, these variations of crop reactions to environmental stresses might be seen in terms of opportunities to increase the diversified farming practices. Here the key idea is to buffer output by maximising diversity in farmers' production systems. This solution is already apparent in the rationale of the adaptation of the existing systems of production in seasonally arid areas.
Strong beliefs in the internal superiority of the two chosen paradigms, and the social border between them, make fertile communication between scientists from these different disciplinary orientations almost impossible. But why should adaptation to drought on the level of the crop and on the level of production system be considered as two separate fields of potential solutions? Why does integration between these different approaches seems to be so fraught with difficulty?

Plant breeders' past experience
Of the 350,000 plant species identified by botanists, 3,000 are known as food suppliers and only 300 are used in agriculture. Looking at the economic importance of the latter, it turns out that the human population lives mainly on 30 species, among which only eight assume a major strategic role in global food security. The greater part of the world's food production is restricted to a handful of species, described by Parlevliet "as not typically adapted to very dry climates".
This is not a strictly technical problem. When breeders have experience in only a few major food crops this is a reflection of world economic history, and shows how the political economy of international food production still influences actual agro­scientific research. International food production is reorganizing and therefore it will be necessary to find out which probably new political­economic relations are now leading change in research programmes, and what the precise influence is of the past trend towards greater reliance of a few uniform crops on the breeding of drought­tolerant plants.
Trend towards greater uniformity of world agricultural production.
The idea of widening the window of climatic opportunity so that 'marginal' environments can be used as much as mid­latitude climatic zones, is quite common. Scientists at Pioneer Hi­bred International are developing drought­tolerant maize and are focusing on processes that occur during flowering ­ a time when maize is most susceptible to drought and subsequent losses in yield. This work is done against a background of work more directly related to genetic engineering. Indeed, Pioneer uses Restriction Fragment Length Polymorphisms (RFLPs) to identify specific drought­tolerance genes. The technique will allow specific DNA bands, correlated with drought tolerance, to be traced to specific chromosomes of parental lines. This is in effect a policy of suiting the crop that helped transform the Mid West of America to better 'fit' the world's varied environments.
A similar emphasis is found in other North American research institutions. For example, scientists at the United States Department of Agriculture are researching the possibilities of widening the thermal window for crops such as wheat and cotton. The thermal window, within which the enzymes of a crop function optimally, for wheat and cotton is between 17.5oC and 23oC. However, for cucumber it is between 30oC and 40^oC. The research goal is to transfer genes encoding key enzymes from a plant like cucumber to plants with lower thermal windows, such as wheat and cotton. The result might be that these transgenic plants will grow at higher temperatures. This will open the tropical world to temperate agriculture.
The point is not to say that this development is necessarily good or bad, but that agricultural research might have different goals if tropical scientists were in a numerical and financial ascendency. Therefore, it will be necessary to find out if biotechnological research on drought tolerance is concentrated on the few major food crops significant in world trade of agricultural products. Another question is whether research institutes of developing countries are swept along by this bend, or will be able to shift their attention towards other crops of more local or regional significance in the attempt to increase the diversity of their agricultural production systems.

Institute's locality
The contents of the breeding programmes of different research institutes are influenced by their social environment. A shift in the distribution of dry periods in US Mid West agriculture might stimulate a growing attention in the USA to develop drought­tolerant varieties. Pioneer's attention to drought­tolerant maize might also be logical, viewing the socio­economic history of the company as a leading representative of USA­based agribusiness. Instead of following the major stream of plant breeding activities on wheat, rice, and maize, the International Crops Research Institute for the Semi­Arid Tropics (ICRISAT) pays major attention of a range of other crops.
In ICRISAT, drought tolerance is a breeding object for sorghum, pearl millet, chickpea and pigeonpea. In fact, sorghum is an important staple in the diets of the people in semi­arid tropics. Worldwide, it is grown on over 40 million hectares. Pearl millet can be successfully cultivated in areas too dry for sorghum and is a good source of fresh and dried fodder for livestock. Finger millet is a food crop in the highlands of Eastern and Southern Africa. Pigeonpea, like chickpea, is a protein­rich crop important in small­scale agriculture in South Asia. ICRISAT's approach of working over a range of different and neglected crops seems to imply a more positive evaluation of existing variations in environments and farming systems.
We will not go into the institutional history that makes an international research centre like ICRISAT so different from private­sector research programmes such as, for example, the programme of Pioneer Hi­bred. The point here is that the dynamic of incorporation of any research group within its regional social environment, will be of increasing importance when choices begin to predominate over necessity.
Local influence, however, might be deflected by or even subsumed within the cultural linkages between researchers at a global level. It might even be the case that these professional linkages weaken the influence of the local geographical environment on a specific research institute  to such extent, that local influences are snuffed out entirely. Case studies might tell us that, for example, biotechnological research institutes in developing countries are heavily acculturated to western­orientated patterns of research behaviour that they are indistinguishable in practice from the research programmes of western private companies. A key issue for social research on abiotic stress will be the extent to which these global, cultural linkages, as distinct from local notions 'set the pace' for the research in developing­country institutes.

It is uncertain how we should weigh the exact influence of each factor. However, it becomes clear that underlying social and cultural ideas can no longer be separated from the breeding activities. Therefore, biotechnological research deserves a sociological approach, since drought­tolerant plants are first and foremost social artefacts.
The neglect of the cultural and social content of drought tolerance research activities is not surprising. We consider it also as a social fact. As George Orwell puts it in his book '1984': "Ignorance is the strength". Indeed, by denying the social base of drought tolerance research, specific social and cultural ideas, materialized in the research results, might be spread over other regions. This would in turn imply that other social and cultural values, which in principle might also be in newly developed products, will become marginalized or even substituted by the dominant ones. Therefore, we emphasize the need to make explicit the cultural assumptions hidden within the paradigms upon drought tolerance research is based.
Guido Ruivenkamp/Paul Richards (Agricultural University Wageningen, The Netherlands)

Sources
Guido Ruivenkamp (1987), 'Social Impact of Biotechnology on Agriculture and Food Processing'. Journal of the Society for International Development, no. 4 ('Seeds of change'), pp. 58­59.

J.E. Parlevliet, A.A. de Haan, and J.J.A.M. Schellekens (1991), Drought Tolerance Research: Possibilities and constraints. Study carried out at the request of the Netherlands Directorate for International Cooperation. March 1991.

International Crops research Institute for the Semi­Arid Tropics (1993), ICRISAT Report 1992. Patancheru, India: ICRISAT.