The Post Office occupational health service was established in 1855. Unique epidemiological records of sickness absence, medical retirements, and deaths for 90 years have been extracted from annual reports. The stability of the death rate is striking, but the male sickness absence series is consistent with three periods of roughly constant rates, viz 7-6, 10-0, and 13-1 calendar days a year with increases in level coinciding with the two world wars. By contrast with general experience, Post Office absence has not shown a rising trend in the past 30 years. An examination of the relation between medical retirement and sickness absence rates shows that a strongly negative correlation has reversed to a strongly positive one since the second world war. The major changes in diagnostic causes of absence, retirements, and deaths are described.