To study accuracy of phase mapping for assessment of the atrial activation sequence during typical atrial flutter, four patients were examined. The Amycard 01K device was used for the study purpose. In accordance with the data of surface ECG mapping, the personal rope geometry, personal heart geometry, and coordinates of electrode for surface mapping, local unipolar electrograms were reconstructed approximately in 2500 sites of the atrial surface. The electroanatomic mapping of the right atrium with activation map construction was made followed by linear ablation of the lower cavotricuspid isthmus with subsequent confirmation of presence of bidirectional conduction block. After termination of the catheter procedure, the data recorded using the CARTO system were exported and processed simultaneously with the data of non-invasive mapping. Three-dimensional atrial models created using the CARTO 3 system were superposed with models created using the tomography data. Non-invasive phase maps showed excitation patterns, which in generally corresponded appropriately to activation patterns obtained during the invasive mapping. Rotation of the activation front clockwise around the tricuspid ring in 3 cases and counterclockwise in 1 case were visualized correctly. In all cases, the onset time of bipolar spikes strongly correlated with phase jumps on unipolar electrograms (r=0.97 0.99 for typical atrial flutter and r=0.99 for atrial flutter with clockwise rotation of excitation). The corrected activation time by the phase jump was 5.7±1.7 ms; the normalized activation time: 2.3±0.5%. Thus, phase mapping permits one to assess correctly the activation sequence during typical atrial flutter with different directions of rotation of the excitation; phase mapping may be considered a promising technique of non-invasive mapping of re-entry arrhythmias.