ABSTRACT  During the Fall Quarter, 1969, at Ohio State University, twenty FL methods students were randomly assigned to two treatment groups—Group I, which received three weeks training off campus, and Group II, which remained on campus—in an attempt to measure the effect of such training. Different instruments were designed to obtain information on items of felt anxiety prior to student teaching, ability to teach Level I Spanish, and various aspects of FL teaching. During student teaching each teacher was videotape-recorded for thirty minutes and his performance rated by two trained observers. Results of the research show that Group I students tend to have less anxiety prior to student teaching, and report higher confidence in their ability to teach Spanish. Performance during student teaching was judged to be nonsignificant. Group I students had significantly higher scores on the ability to teach the dialog; their scores were also significant on the variables of how dialogs, vocabulary teaching, pattern practice, and grammar generalization “fit” into an audiolingual program. Ability to adapt or supplement materials for teaching seems to be unaffected by the direct experiences. Guidelines for implementing the described experiences were also given.