The use of a modified training opportunities approach for predicting the number of graduate enrollments and degrees at the University of Illinois at Urbana-Champaign is considered. Data concerning graduate assistantships and faculty staffing levels were used to predict, with a high level of accuracy, the number of graduate enrollments and degrees. Predictions were based on the multiple regression technique. Possible impacts upon institutional, regional, and national policy are discussed, as are the other major theories of graduate manpower planning: the market, demographic, and credentialist approaches. From a theoretical perspective the study tends to support the notion that the training opportunities approach can be used as a proxy for the demographic, market, and credentialist approaches in analyzing scientific manpower supply and demand. The basic assumption of the approach was that the availability of financial aid to students in scientific fields, as well as the faculty staffing levels in those same fields, ought to serve as a useful predictor of enrollments and degrees produced. (Author/SW)