A study of myocardial Mg, based on 1236 samples obtained at post-mortem examinations, is reported. There was a lower Mg level in samples from subjects certified to have died from IHD causes, by about 13%, than in samples from other deaths. There were lower levels, by about 5% where death had been lingering, compared to sudden, and there was a large range, equivalent to +/- 12%, in the mean Mg levels in samples from different areas. Various possible sources of bias are briefly considered.