A performance comparison of the following five leads is made in the homogeneous torso model: The sagittal leads of the Frank, Schmitt-Simonson SVEC-III, and McFee-Parungao lead systems, a simple bipolar lead consisting of a precordial and dorsal electrode, and a 7-by-5 parallel grid sagittal lead. The criteria of uniformity, orthogonality, and ability to measure the total outward dipole moment of the heart are used in judging lead quality. A point of reference corresponding to the approximate center of gravity of the ventricles, and concentric “large”, “normal”, and “reduced” heart areas, are defined within the model and on maps placed on a drawing board near the model. A current source able to deliver x, y, and z-directed current, is three directionally movable within the model and its position is transmitted to the drawing board by means of a rigid arm. The source, z directed at the point of reference within the torso, is excited to deliver current which is sufficient to cause a 100-mv. reference voltage in the lead under study. Lead voltages obtained when the source is x, y, or z directed at 177 points within the cardiac area of the torso, are then read directly as percentages of the reference voltage. In an ideally uniform lead, 177 readings of 100 mV. are obtained when the source is z directed. In an ideally orthogonal lead, two sets of 177 zero readings are obtained when the source is x and y directed, respectively. 1. 1. By inspection alone of graphs which show the frequency distribution of individual voltage readings around these standards of ideal accuracy, and of torso maps which show lines of percentage deviation from the reference voltage, leads are grouped as follows: (i) the least accurate group, containing the Frank, McFee-Parungao, and bipolar sagittal leads; (ii) the SVEC-III sagittal lead, intermediate between the first and third classes; and (iii) the grid, showing greater uniformity and orthogonality than the four other leads. 2. 2. Nonuniformity expressed as arithmetic standard deviation of measured lead voltages supports the division into three classes. It is 30 to 50 per cent in group (i), 16 to 24 per cent in the SVEC-III sagittal lead, and 4 to 9 per cent in the sagittal grid. The standard deviation on a logarithmic scale of millivolts again separates the three groups (p = 0.001), and within the least accurate group (i) there is, in most of the sets analyzed, no significant difference in uniformity between the Frank, McFee-Parungao, and bipolar leads. 3. 3. The means of sagittal lead voltages when the current source is x or y directed, respectively, used as a measure of nonorthogonality, are 12 to 25 per cent in the Frank, McFee-Parungao, and bipolar group, 12 to 30 per cent in SVEC-III, and 4.3 to 5.7 per cent in the grid. The differences between the grid, on the one hand, and the 4 other leads, on the other, are significant (p = 0.001) in all heart sizes. Differences in orthogonality between leads other than the grid are statistically not significant in most of the samples. 4. 4. The equation M z = A z V z ϱ , derived previously in this laboratory, is accurate within an error of ±5 per cent in measuring the dipole moment of z-directed current source in the homogeneous torso model, if A z is the area of the 7-by-5 grid, V z is the sagittal lead voltage, and ϱ is the torso resistivity. By the performance standards used in this study, the sagittal leads of the Frank and McFee-Parungao systems are not more uniform or orthogonal than is a simple bipolar sagittal lead. The SVEC-III sagittal lead is more uniform but not more orthogonal. The parallel grid sagittal lead is more uniform and more orthogonal than any of the other leads tested.