The extent of variety in structure and format of teaching mathematics in fifth grade classrooms and its relation to student attention was studied. The data for this study consisted of observations made in eighteen math classes for a period of six to ten days with a mean of over eight days. Seatwork and recitation accounted for over 95 percent of student-minutes, which indicated that little variety existed in mathematics instruction in the observed math classes. Socioeconomic status and level of expenditure were related to attention span, but did not influence the mean number of instructional segments per day. It was concluded that, in teaching a skills-oriented course to fifth grade children, variety of instructional procedures is not an important variable in accounting for student attention. Variety of procedure enhances student attention only if there is sufficient time in the class period to allow for more segments. If there is not, it detracts from attention. It was observed that on days with more segments than usual, attention increased somewhat, and a less frequently-used instructional format brought higher attention levels than the class' more frequently-used format. (JD)