Content and language integrated learning (CLIL) has been examined in education internationally, and related curricula have been developed by many schools in Japan. The current study sought to clarify the effects of CLIL in music ensembles in secondary education on the musical performance of students, and to examine how music teachers can improve their method of instruction. Ensemble activities for small groups were conducted in the ninth grade, in English. Students’ learning process was analyzed in terms of activities, final performance, and self-review, focusing on the following features: (1) development of musical skills, (2) improvement in the method of expression, and (3) group work in cooperation. When performing activities, students worked together more cooperatively compared with before CLIL. This result suggests that tendency for students to help and encourage each other is an important factor in CLIL music lessons. In the last phase in which the lessons were summarized, students who were not proficient in English tended to miss complicated ideas or feelings addressed by the teacher. This finding suggests that teachers should use English and Japanese differently from one situation to another, particularly in the early stages of introducing CLIL into music lessons.