The 1987 Preliminary Report of the international maize testing program covers trials sent out in 1986 and for which data were received from December 16, 1987, to April 15, 1988; it also includes addenda reporting data that were received late for 1985 and 1986 trials. 1987 trial data that are received after April 15, 1988, will be included in the 1987 Final Report. These reports are the end product of carefully conducted trials and the timely return of complete and accurate data by hundreds of trial cooperators, whose participation is extremely valuable in the development of superior germplasm for maize farmers in developing countries. We at CIMMYT wish to thank the scientists that are participating in the international testing program for their time and effort, and we look forward to continued cooperation with them. Population Improvement at CIMMYT A large part of the maize research at CIMMYT is directed toward the development and improvement of gene pools and populations that offer high yield potential and yield stability under a wide variety of production conditions in developing countries. The international testing program serves as a vehicle for distribution and evaluation of germplasm developed by the various Maize Program lJf1fts and thus links them with cooperators throughout the Third World. In its population improvement work, the Program has employed the modified full-sib recurrent selection procedure illustrated in Figure A. With this procedure one cycle of improvement can be completed in four cropping seasons or two years. Trials are grown both north and south of the equator. Since the main growing season occurs at different times in the two hemispheres, it takes two seasons for all of the trials to be grown in both hemispheres and the data collected. Maize Program breeders make good use of these extra seasons by carrying out within-family improvement for the traits that are most important in particular populations and by Including one recombination generation for selected families.