One hundred and forty secondary teacher candidates who were enrolled in a first course in teaching were the subjects of a seven-week study to determine the effects of microteaching in a Teaching Laboratory (TL) on verbal behaviors. All students taught a 10-minute pretest prior to the start of the study and then were divided into two groups. Group A students (85) taught five- to eight-minute TL lessons with feedback and subsequent reteach lessons in two cycles (emphasizing clarification of instructional objectives and interaction). Group B students (55) had no direct teaching experience. At the end of the seven-week period, all students taught a 10-minute posttest. When both sessions were measured using the Laboratory Observation Schedule and Record (LOScAR) and when data obtained were subjected to analysis of covariance, results indicated that teaching candidates' verbal behaviors can be modified in a TL. Adjusted group means showed statistically significant differences in 17 of the 22 variables of the LOScAR favoring group A. (A 16-item bibliography is included.) (SM)