Under certain circumstances it is possible to record His potentials using surface chest leads. These potentials have a very low intensity over the thorax, and must be separated from the "background noise" by an averaging technique. The desired activity occurs before the R wave and has a constant relationship to it; the potentials picked up by the chest leads are amplified, coded numerically, and presented to the memory store which accepts only those potentials which precede the R wave. It then summates the successive cycles, using a synchronising signal to ensure that they coincide. All this is carried out with an averager. In the 80 patients studied, a signal which seems very likely to reflect His activity was obtained in slightly more than a third of cases. The findings were correlated in 14 cases with those produced by invasive techniques, and showed that the main source of error was the after-potential of the P wave after a short PR interval--this occurred in one case; in the other 13 cases, the correlation was good. This method needs to be refined before it can be used clinically. It seems unlikely that it will replace the invasive technique which is so important for stimulation tests, and so necessary for studies of dynamic pharmacology.