In order to investigate the contribution of medical care to the widening of mortality differences between socio-economic groups, we studied differences in the decline of mortality from conditions which have become amenable to medical intervention. For England and Wales, data on occupational mortality from the Decennial Supplements of the years around 1931, 1961 and 1981 were used. For the Netherlands, a more indirect approach had to be followed, using data on geographical variation in mortality for the years around 1952, 1962, 1972 and 1982. In England and Wales during the period 1931-1961 mortality from conditions which became amenable to medical intervention generally declined more in relative terms among the higher occupational classes, both for men and for married women. During the period 1961-1981 the picture was less consistent, although for about half of the conditions mortality declines were again larger in the higher occupational classes. The findings for the Netherlands suggest a similar pattern of differential mortality decline for a small number of conditions only. Possible explanations for these findings are discussed. We conclude that: the widening of the mortality differences between socio-economic groups in England and Wales was partly due to differences in decline of mortality from conditions amenable to medical intervention; the same may be true to some extent for the Netherlands, but the evidence is not striking; differences in decline of mortality from 'amenable' conditions were probably due, at least in part and at least up to 1960, to differences between socio-economic groups in accessibility, utilization or quality of medical care.(ABSTRACT TRUNCATED AT 250 WORDS)