of the background of those numbers and of the organization's plans for continuing diversification. Because there are so many differences in the circumstances of our grantees, we have shunned rigid rules. Instead, we encourage the staff to work closely and judiciously with grantees to define realistic goals for each grant period and to plan an approach to address recruitment or selection problems. Recently, in a number of competitive grant selection processes, we have made diversity in the applicant organization a basic criterion for participation in the competition. We have noted an improvement over time in the affirmative action profiles of many grantees, although few are entirely satisfied with their achievements to date. The progress in numerical representation has been accompanied by changes in attitudes as well. We are heartened by an increased recognition of the value of diversification and by a willingness to promote that objective.

Various aspects of our external affirmative action policy and practice have evolved over the past thirteen years, in part as a result of reviews we conducted in 1975, 1977, and 1983. Our experience suggests that we need to approach this subject early in grant negotiations; consider each grantee's situation individually, often discussing recruitment and selection procedures; and attend to such specific questions as staff or board turnover projected during a grant period.

The most recent reexamination of our experience with affirmative action policy included interviews conducted by a consultant with sixty Foundation staff in the United States and developing countries, and with more than 100 members of grantee organizations in this country and abroad. The findings highlighted the need to update the Foundation's statement describing our policy to take account of knowledge gained since 1972, to initiate ongoing discussions with staff and grantees on the rationale and practice of the policy, to give attention to ways of measuring progress in groups of grantees over several years (perhaps through data storage and retrieval now simplified by computers), to work with our Trustees and with other funders who may be engaged in similar efforts, to make more grants that explore the theory and practice of affirmative action and similar programs around the world, and to call attention to the many positive changes that have occurred in grantee organizations since the policy was instituted. We are now working on each of these recommendations.