Data on monthly totals of cardiac deaths in England and Wales were examined in different ways to see whether there were any unexplained fluctuations in rates both at the time of an unplanned increase in the bioavailability and therefore the potency of the Lanoxin brand of digoxin in May 1972 and also when there was a coordinated increase in the bioavailability of other brands in October 1975. Despite advice to prescribers, dosages were not proportionately reduced in the 600 000 patients who were on treatment. Monthly totals of deaths from cardiac causes were high through the summer of 1972 but not at the end of 1975 and the excess in 1972 seems to be related to a cold summer. There was no evidence of a consistent or specific effect of changes in digoxin potency, either beneficial or harmful, on deaths from all cardiac causes or in specific subgroups in which digoxin treatment was likely to be most common. Although major changes in digoxin potency in England and Wales did not seem to produce a repeatable effect on death rates, data from other countries should also be examined for evidence of such an effect.