Addressing shortcomings of the self-report Toronto Alexithymia Scale (TAS), two studies were conducted to reconstruct the item domain of the scale. The first study resulted in the development of a new twenty-item version of the scale--the TAS-20. The TAS-20 demonstrated good internal consistency and test-retest reliability, and a three-factor structure theoretically congruent with the alexithymia construct. The stability and replicability of this three-factor structure were demonstrated in the second study with both clinical and nonclinical populations by the use of confirmatory factor analysis.