Peer evaluation was implemented in a college composition seminar to explore its effectiveness. The class of 13 students met once a week, workshop style, for 4 hours. Students wrote short papers (three to five typed pages) every week in an assigned mode (critical analysis, social commentary, familiar essay) but on topics of their own choosing. For the first assignment, students and the instructor responded by editing all the essays line by line. The next step was primary trait scoring. Other evaluation methods used included the nonjudgmental description of the essay suggested by Peter Elbow, analytical scoring moving from problems with the essay as a whole down to grammar and mechanics and even transactional analysis. The class discussed audience expectations, genre, and levels of style. Such discussions, outside as well as in class, led to thoughtful revisions. The writing of all the students improved, due to total immersion in the writing process and increased audience awareness. Course evaluations indicated that students also perceived an improvement in their own writing. Editing became more incisive and reliable as well. At the end of the term students graded all their peers on the quality of both their editing and their writing. The students' evaluations did not differ from those of the instructor. Despite the work load, students spoke quite fervently about the value of the course, indicating that peer evaluation can be a very effective teaching tool. (HTH)