Although the determinants of radiofrequency lesion size have been characterized in vitro and in ventricular tissue in situ, the effects of catheter tip length and endocardial surface orientation on lesion generation in atrial tissue have not been studied. Therefore, the dimensions of radiofrequency lesions produced with 4-, 6-, 8-, 10-, and 12-mm distal electrode lengths were characterized in 26 closed-chested dogs. The impact of parallel versus perpendicular catheter tip/endocardial surface orientation, established by biplane fluoroscopy and/or intracardiac echocardiography, on lesion dimensions was also assessed. Radiofrequency voltage was titrated to maintain a steady catheter tip temperature of 75 degrees C for 60 seconds. With a perpendicular catheter tip/tissue orientation, the lesion area increased from 29 +/- 7 mm2 with a 4-mm tip to 42 +/- 12 mm2 with the 10-mm tip, but decreased to 29 +/- 8 mm2 with ablation via a 12-mm tip. With a parallel distal tip/endocardial surface orientation, lesion areas were significantly greater: 54 +/- 22 mm2 with a 4-mm tip, 96 +/- 28 mm2 with a 10-mm tip and 68 +/- 24 mm2 with a 12-mm tip (all P < 0.001 vs perpendicular orientation). Lesion lengths and apparent volumes were larger with parallel, compared to perpendicular tip/tissue orientations, although lesion depth was independent of catheter tip length with both catheter tip/tissue orientations. Electrode edge effects were not observed with any tip length. Direct visualization using intracardiac ultrasound guidance was subjectively helpful in insuring an appropriate catheter tip/tissue interface needed to maximize lesion size. Although atrial lesion size is critically dependent on catheter tip length, it is more influenced by the catheter orientation to the endocardial surface. This information may also be helpful in designing electrode arrays for the creation of continuous linear lesions for the elimination of complex atrial tachyarrhythmias.