This is an evaluation of a project whose expressed purpose is to teach "abstract conceptually-oriented" mathematics to disadvantaged elementary school children using a discovery-oriented, pattern-recognition approach. This report is primarily of Project SEED (Special Elementary Education for the Disadvantaged) in Gary, Indiana during 1970-72. It attempts to evaluate the affective impact on children and the impact on the teachers' behavior; instrument construction in support of these evaluation areas is discussed. General conclusions are that students with SEED instruction do somewhat better in mathematics and that most like the program. There is no clear-cut evidence that this instruction has a significant effect on the child's general self-concept or that it has a lasting impact on the regular teachers' behavior. A serious problem remains concerning educational priorities and cost effectiveness of the project. (LS)