Myringotomies were performed on 75 ears suspected of having chronic or recurrent otitis media with effusion. Preanesthesia and intraoperative tympanograms with halothane, nitrous oxide, and oxygen anesthesia were compared for possible changes in middle ear pressure due to anesthesia. In addition, preanesthesia tympanograms were compared with tympanograms taken 2 weeks before hospital outpatient surgery. Tympanograms of 11 ears were found to shift from a flat type recorded at the office visit to normal range at pre and postinduction anesthesia. Middle ear fluid was not present in any of these 11 ears. The remaining 64 ears were broken down into three abnormal tympanogram configurations; flat, roll-over, and peaked negative pressure types. Comparing the tympanograms done in the office, preanesthesia and intraoperatively, only one of the 64 ears showed a change after induction of anesthesia. Three of the 64 ears revealed an unpredictable result with negative myringotomies. These results tend to support tympanometry as a means in predicting myringotomy findings, and also refute the contention that short-term nitrous oxide anesthesia changes the middle ear effusion during myringotomy.