Objectives: The purpose of this long-term study was to evaluate angular and linear measurements of cephalograms and determine where the vertical relapse occurs after deep overbite correction. Methods: Forty-three patients were evaluated using three laterally oriented cephalometric roentgenograms. Roentgenograms were taken before orthodontic treatment (T1), at the completion of treatment (T2), and at least 5 years after the removal of all fixed appliances. All cases in this sample were treated with a Straight-Wire Appliance (0.022-inch slot appliance - twin brackets) with continuous wires with a reverse curve of Spee and/or an accentuated curve in the upper arch. The retention protocol was similar in all cases. Cephalograms were analyzed for each patient at T1, T2 and T3. Serial cephalograms were hand traced and analyzed by a digitizing tablet (Numonics, Landsdale Pa) and digitizing software (DFP Plus 2.02, Dentofacial Software, Toronto, Ontario, Canada). Sixty-eight landmarks were located and digitized and thirty-eight variables were analyzed in the study. The study group was compared with a similarly matched group using normal growth data (Michigan Growth Study, 1974). Statistical testing was done by performing Analysis of Variance, T-Tests, and Pearson's Correlation with the level of significance (alpha) set at 0.05. Results: Analysis of the data from each of the forty-three patients showed a mean overbite reduction of 3.8 mm and a mean relapse of 1.6 mm. During treatment, the lower molars significantly extruded, which accounted for most of the overbite correction. The lower incisor angulation changes during the post-treatment period demonstrated a high correlation with the relapse of the deep overbite correction. Conclusions: It was concluded that deep overbite correction by extrusion of molars is stable when there was compensatory ramus growth and it is the post-treatment lower incisor labio-lingual changes that are significantly correlated with the lack of stability of the correction.