The purpose of this experimental study was to determine whether instruction in generative grammar would produce better student writing. As a secondary aspect of the study, evaluation of all students was made for growth in knowledge of grammar apart from writing or other practical applications, whether the students had been instructed in generative grammar or traditional grammar. For evaluation, advanced experimental students were compared with advanced control students, and low experimental students were compared with low control students. Students were administered Form 3B of the STEP Writing Test at the outset of the experiment and Form 3A of the test at the conclusion. A teacher-made, objective-type grammar test in two forms was used to measure growth of the students in knowledge of grammar. Findings show that instruction in either generative or traditional grammar would prove equally effective with advanced students in bringing about growth in knowledge of grammar; and that instruction in generative grammar would prove more effective than traditional grammar in knowledge of grammar in low students. Whether generative or traditional, instruction in grammar did not significantly improve student writing ability in either advanced or low students. (Author/DB)