1. 1. Total QRS duration, from earliest to latest deflections of ventricular activation (depolarization), is not measurable in limb leads, due to isoelectric projection of early or late parts of the spatial QRS vector loop. 2. 2. The most nearly correct value for QRS duration can be obtained only from simultaneously recorded, orthogonal XYZ lead derivations with high fidelity, high-speed recording and tedious measurement technique. 3. 3. Total QRS duration in XYZ leads was measured in 115 men, aged 20 to 50 years. The range of values was 0.0801 second to 0.1309 second, the mean 0.1010 second, the standard deviation 0.010 second. 4. 4. It is suggested that 0.12 second be the "upper normal limit" of QRS duration in XYZ leads. Fifty-four per cent of the values were over 0.10 second; 18 per cent over 0.11 second; and 3.5 per cent over 0.12 second. 5. 5. V1 and V2 are suggested as the single leads in conventional records in which the more nearly correct QRS duration may be most often measured. 6. 6. The QRS duration found in isochronous recordings of Leads I, II, and III was uniformly shorter than that in XYZ leads and did not permit prediction of the spatially correct QRS duration. The range of QRS duration in simultaneous records of Leads I, II, and III was 0.765 to 0.109 second, the mean 0.904 second. 7. 7. Poor correlation of QRS duration with heart rate was found except at the extremes of heart rate where correlation was high. 8. 8. Poor correlation was found between QRS duration and body weight (heart size implied). 9. 9. QRS duration increased slightly after exercise in the face of accelerated heart rate. 10. 10. Mean beat-to-beat change in total QRS duration from XYZ leads was on the order of 3 msec., while beat-to-beat change in single lead strips averaged 2 msec. 11. 11. Separation of all positional and other factors in beat-to-beat variability of QRS duration awaits similar measurement in three-channel simultaneous records from apparatus of comparable sensitivity. 12. 12. Subtle beat-to-beat changes, QRS duration change with exercise in normal and abnormal subjects, and "spatially corrected" P-R and Q-T intervals are obviously indicated studies.