With the immediate aim of reinforcing the productive and receptive control of ordinary spoken language, experiments were undertaken using two methods of teaching French to university students with some knowledge of the written language. The first of the two groups involved in the experiment completed a 240-hour course consisting primarily of dialog repetition. The 200-hour course attended by the second group included, in addition to the dialog repetition, much partially programed morpho-syntactic pattern drill in the language laboratory. Programing considerations were the phonetic features (particularly of the Sandhi phenomena) that are the main sources of error in decoding spoken French. The series of tests administered at the beginning and end of the courses indicated that both groups had achieved the same degree of improvement. However, the fact that the second group had accomplished the improved level with 40 fewer hours showed that dialog repetition, used in conjunction with pattern drills, was the more effective method. (JH)