A partial immersion bilingual (English-Ukrainian) program which included Grades 1 to 3 was examined to determine: (1) the program students' achievement in mathematics, English and Ijrainian; (2) the attitudes of parents and program personnel toward the program; (3) the program students' level of cognitive development; and (4) reasons students left the program. Six tests of cognitive and linguistic development were individually administered to 12 children in regular classes, 12 fluently bilingual program children and 12 nonfluent program children in grades 1 and 3. The performance of the nonfluent bilingual program first and third grade children was similar to that of the regular students on five of the tasks. On the sixth, the Semantic-Phonetic Preference test, the first grade nonfluent bilingual children made fewer semantic choices. The fluently bilingual children in both grades 1 and 3 were better able to analyze ambiguities in sentence structure than either the nonfluent bilinguals or the regular students. On the other five tasks, there was no significant difference between the fluent bilinguals and the other two groups. In general, the majority of the teachers and administrators thought the program objectives were appropriate, considered the program a success, and wished to see it continue. Parent attitudes were also positive. (Author/JAC)