The direct-current magnetocardiogram not only shows the features usually seen on the electrocardiogram, but also shows the T-Q (baseline) shift due to cardiac injury current. The first direct-current magnetocardiogram measurements of the human heart are presented here. The hypothesis tested is that there is no injury current associated with the S-T shift seen in two electrocardiogram abnormalities: early repolarization, and left bundle branch block. The data from three typical early repolarization subjects and one typical left bundle branch block patients are presented. It is found, in each case, that although there is appreciable S-T shift, there is essentially no baseline shift on the direct-current magnetocardiogram. The absence of baseline shift proves that the S-T shifts in these cases are not 'apparent' shifts, caused by a dc injury current which is interrupted during the S-T interval; instead, these are 'true' S-T shifts caused by a current flowing only during systole, presumably due to an altered repolarization of the ventricles. It is also found that the direct-current magnetocardiogram does not have routine clinical application because of a practical problem. This is the presence of false baseline shifts due to noncardiac currents, mostly in the gastrointestinal tract, which could be suppressed in only about one-third of the subjects. However, the direct-current magnetocardiogram may be useful as a research tool, for clarifying the cause of the S-T shift in selected subjects.