Distress tolerance (DT), or the ability to withstand psychological distress, has been proposed as a mechanism underlying multiple forms of psychopathology. However, research on DT is limited in several areas. First, stability and change of DT over time has never been assessed in adults. Second, it is unclear whether alternative conceptualizations of DT yield differences in longitudinal stability and change. Third, gender differences in DT have yet to be examined in nonclinical adult samples. And fourth, longitudinal predictive utility of DT has not been adequately assessed. The purpose of this study was to investigate these 3 questions using data collected at 3 time points over a 6-month period, examining borderline personality disorder (BPD) features as an outcome. Using 3 different measures of DT, results indicated that there is no mean level change in DT. Similarly, there was moderate rank-order stability in DT and no significant individual level change across measures. These findings suggest that DT is similar to other stable, trait-like constructs, as has been previously theorized. Next, a series of cross-lagged panel models revealed that although DT had a cross-sectional relationship with BPD features across all time points, DT did not predict BPD traits longitudinally. These findings have implications for treatments for BPD.