The architects of this session have highlighted an important theme in a timely manner, although I have yet to discover when there was an era in which resources were not scarce. Part of the timeliness relates to the links between these papers and the imminent production of three important monographs: one on methods of priority setting that I will refer to here as (Norton, Pardey, and Alston); the second reports a major exercise being undertaken by TAC/CGIAR that I will refer to as (McCalla and Ryan); and the third an in-depth analysis of the U.S. agricultural research system (Huffman and Evenson) that I will refer to as (Just and Huffman). Constraints of time and space being what they are, the authors have done well to do as much justice as they have to the scope of their respective fuller works in these presentations. Analogous constraints apply to the discussants and, accordingly, I will focus on just a few key features of these welcome and commendable contributions. Most of my remarks are addressed to the first two papers, for the good reason that they arrived in time to be read before these comments were due, but Direction (and the related book) is clearly an extremely useful contribution because of the size, stature, and global significance of the U.S. system (especially as a much-albeit imperfectly-mimicked role model), and because of the course it seems to be headed. In the preliminary version of Direction that I had the chance to skim, I particularly appreciated the analysis of why the advantages of the (oft-vaunted) competitive grants in publicly funded research are so exaggerated. One common issue in the two papers, which are so centrally addressed to tradeoffs, is yet another (I think rather under-mentioned) tradeoff between transparency and comprehensive realism. The latter attribute in turn determines credibility, and just whose credibility is to be influenced depends naturally on their perceptions of what features of reality (and the economic concepts used for capturing it) are most important to embody in a model. It seems that Lessons have opted for many extreme simplifications in their desire to have a simple, transparent, and readily computable version in spreadsheet form whereas Issues, as is probably inevitably the case, have taken a generally more vague but broader and less-confining approach to this issue. One gathers, however, from the applications cited in Issues, when it comes down to the practicalities of attempting concrete projects in this field, again they, perhaps unsurprisingly, opt for simplicity and pragmatic transparency in lieu of nonlinearities and other niceties (Anderson). My diffidence about throwing stones too big is traceable to the concerns of one who lives in a thinly glazed house. I have been somewhat critical of several aspects of these papers on earlier occasions (e.g., Anderson) and for brevity I will not repeat all my continuing concerns. My anxiety is that, were I pressed into empirical service for priority setting of the types discussed here, I am confident that I would be at least as guilty and thus I have to congratulate those connected with Issues and Lessons for their courage in taking on work that is so readily the target of attacks of many kinds yet, as they variously argue, must indeed be done by someone if the complex decision analysis in these processes is to be improved. Institutions that supply the resources to increasingly hard-pressed public research system (at both international and national levels) will need more formal structures of the types overviewed here to feed their own bureaucratic requirements for accountability and scrutiny. While the issues have been clearly identified in the discussion of the size of the supply shift, Jock R. Anderson is a principal economist, Agricultural Policies Division, Agriculture and Rural Development Department of the World Bank.