The tropical AKST institutions established by the colonial powers, such as the Royal Tropical Institute (Netherlands) or the Institut de Recherche pour le Développement (formerly ORSTOM) (France) and their supporting university networks similarly have surrendered their dominance and undergone major transformations over the period (Jiggins and Poulter, 2007), yet they remain collectively the largest source of knowledge on the diversity of ecological and ethnic situations in the tropics. These institutional arrangements were generally effective for their initial purpose, but they badly neglected the food crops consumed by indigenous populations, with the exception of a few such as the federal research station for French West Africa created in 1935 to increase food production (Benoît-Cattin, 1991). The International Agricultural Research Centers (IARCs), subsequently grouped under the CGIAR umbrella, was in part a response to this gap.

CGIAR. Assessing the role of the CGIAR is fraught with difficulties, mainly because of the controversies raised by this important actor, since its inception. Several external reviews of the CGIAR took place in the 1990s (World Bank, 2003), most of them organized by the CGIAR itself, indicating a willingness to change and adapt but also some uneasiness about the way the CGIAR worked, chose its priorities and was governed. However, the reviews did not fully address some of the more fundamental questions raised by the critics. There is insufficient space here to do justice to all these debates.

Creation of the CGIAR. The role of the two US-based foundations, Rockefeller and Ford, in the creation of the first international centers has been well-documented (Baum, 1986). The first international research center, the International Maize and Wheat Improvement Center-in Spanish, CIMMYT-was devoted to wheat and maize, the second one-the International Rice Research Institute (IRRI) established in the Philippines in 1960-to rice. This early emphasis on cereals, i.e., on staple food crops, was a direct reaction, befitting the philanthropic nature of the two foundations, to the emphasis on plantation crops during the colonial era.

     The emergence of this new type of institutional configuration had a profound impact on the IAs for agricultural research in developing countries. In this respect, the rapid evolution of the role of IRRI is exemplary. The first highyielding (HY) rice cultivar released by IRRI (IR8) grew out of a dwarf gene which originated in Japan. Soon, however, its limitations became obvious. The new variety was sensitive to multiple pests and did not have the taste desired by many in Asia. The second generation of HY cultivars released by IRRI grew out of elaborate collaborations among many national research institutions in Asia, permitting a quantum jump in the exchange of genetic material and the coordinated testing of new genetic material in multiple locations. These new kinds of IA's, based on networking among public research institutions, with the hub located at an international center, set a pattern for the future. The role of the international centers in the development of new and more productive staple crop varieties has been well documented (Dalrymple, 1986) and is not by itself a controversial issue.

Early criticisms. But early on, extensive criticisms were expressed; in particular, it was pointed out that a technological change, however rapid and even if called a revolution as in the expression "green revolution", could fall short of the radical changes in agrarian structures which many felt were necessary to tackle the most glaring inequalities associated with unequal access to productive resources, land in particular (Griffin, 1979; Griffin and Khan, 1998). One must recall in this respect that the Green Revolution (GR) came after many attempts at promoting land reforms or agrarian reforms. Many of the reform attempts were made in a climate of bitter social struggles, often violent. In this context, the promotion of an international consensus in support of a technology-led green revolution could be seen as an alignment with conservative forces, nationally and internationally (Frankel, 1971). A similar criticism saw the GR, the CGIAR and their promoters such as the World Bank, which indeed had played a crucial role in the formal creation of the CGIAR, patterned on other consultative groups sponsored by the Bank, as an attempt at "liquidating peasantries" in developing countries (Feder, 1976). These criticisms prompted a large body of empirical research and interpretative analyses to evaluate the impact of the GR on poverty and the survival of small-scale producers. The assessment of the merits and limitations of the transfer of technology (ToT) model draws on that literature (2.1) (Harris, 1977; Lipton and Longhurst, 1989; Biggs and Farrington, 1991; Hazell and Ramaswamy, 1991; Lipton, 2005). One important lesson was that the social impacts of the technological changes associated with the GR varied greatly in space and in time. This should not come as a surprise since we know that technological change is only one component in the complex evolution of social realities yet the implications for how AKST were conducted within the CGIAR and with the CG's partners did not immediately sink in. The controversies themselves also reflect the fact that many views expressed in the controversies were oversimplifications drawn from limited empirical data, giving privileged attention to some aspects of the complex phenomena involved.

     Similarly, the debates on the role of the CGIAR in the impact of the GR on the environment have been heated (2.1). Those who defend the GR and the CGIAR emphasize the millions of hectares of primary forests and other lands saved from destruction through the intensification of existing cropland that the GR permitted (see Borlaug's numerous public speeches on the topic). There is no doubt, however, that negative environmental effects, ranging from pollution to degradation of land and water resources also have been significant (Byerlee, 1992). Another environmental consequence, the increase in the uniformity of crop germplasm, with all the risks that the corresponding loss of biodiversity entails, roused similar controversies (Hogg, 2000; Falcon and Fowler, 2002).

Subsequent evolution of the CGIAR. These debates and the recognition that many issues were not well addressed led to changes within the CGIAR. For instance, it was recognized that the focus on individual crops had serious limitations. Mixed farming-the basis of many small-scale farming systems- agroecosystem sustainability and the management of natural resources had not been addressed systematically.