The effects of two different types of bilingual science instructors upon student science achievement in a bilingual science program were investigated. The influence of student bilingual dominance upon science criterion scores were used to determine program success and teacher effectiveness. Sixteen selected junior high schools in New York City with large percentages of Hispanic-background students were selected for the study. In each school there was one seventh grade class taught in Spanish and English, and one control class taught in English. Nine schools had fluent Spanish speaking science teachers; the seven other schools had science teachers familiar with, but not fluent in, Spanish. Students were randomly assigned to experimental and control classes and were tested for initial science ability and bilingual dominance, and were post-tested for science achievement. Results indicated that bilingual classes taught by teachers fluent in Spanish made significantly better achievement gains than bilingual classes taught by non-fluent teachers. (BC)