At the National Endowment for the Humanities the greatest concentration of applications involving the computer is in the Research Tools Program within the Division of Research Grants.1 It considers about 220 applications each year of which 75 employ a computer in some way. This number represents approximately 34 percent of all applications submitted to the tools program.2 The chief reason for this clustering is that the research division was originally established to foster scholarship that required more wherewithal than the Division of Fellowships could provide through salary maintenance and modest travel allowances. The Division was also seen as facilitating collaborative research, longrange projects, and ones of unusual complexity.3 Unlike the fellowships division, it is empowered to award grants of a rather considerable amount to cover not only the costs of salary, even when several persons are working on a project, but also travel and per diem expenses, supplies and rental of equipment, honoria for consultants, and, where appropriate, the preparation of printer's copy in "camera-ready" form. Understandably, then, persons with ambitious research undertakings frequently come to this Division to seek funding, especially when a computer and related costs are involved. The evaluation of these proposals engages a proportionally larger amount of time within the Division because of the often technical nature of the proposals and because of the sizeable amounts of money involved. In the following, I should like to discuss the history of a rather typical, if small, group of projects relating to French language and literature funded in previous years by the Division and then to describe the review process that has evolved from this experience. One of the first applications using a computer to be funded by the Division, before the creation of a separate tools program, was Professor Jerome Clubb's "Automation of the Statistical Sources of French History: the Statistique Gindral de la France." Although this project has nothing to do directly with French literature or language, it is instructive to note a few of its features to see what changes have occurred since 1971 in terms of technology and methodology.In addition, these changes will assume a rationality since the review process is a dynamic one, effected by and effecting the types of proposals it considers. To begin with, Professor Clubb encoded his data on more than one million punched cards-a procedure that today would hardly be contemplated. Secondly, the data base thus created could provide only limited service in the form of simple analyses and sorts of the accumulated material. In a narrative report on this grant, Clubb acknowledged that punched cards were awkward and inefficient and that the computer was being used primarily to process material-that is to match, verify, and adjust for inconsistencies-rather than to perform elaborate sorts or to retrieve information. This suggests, of course, that the data-base aspect of the project was less successful than anticipated. In fact, we in the Research Tools Program have no specific information on the use that has been made thus far of this research material. Obvi-