McKeown and Record's classification of the causes of the nineteenth- century mortality decline has proved a fruitful one and forms a useful framework for considering the distinctiveness of the demographic findings of Preston and Haines. 1 McKeown and Record detected four main controls on mortality: medical science, the standard of living, the environment, and the virulence of disease organisms. They argued that changes in the mortality from particular causes of death could be assigned to the operation of one or other of these main controls. For England and Wales, they argued that medical science had little to offer in fighting any of the main diseases that actually waned, except for smallpox, which was attacked through vaccination. Thus the small contribution of smallpox to the mortality decline registered the slight contribution that the medical profession had made to improving individuals' life chances during the second half of the nineteenth century. Believing that diet (and thus the standard of living) controlled the level of mortality from tuberculosis, they proposed that the large contribution of the decline in tuberculosis mortality to the overall mortality decline was clear evidence that the main contribution to English and Welsh longevity came from a general rise in real incomes. The intermediary y position of the diarrheal diseases placed environmental changes (the sanitary revolution) above medicine but below economic growth. Finally, a small residual contribution from some childhood diseases whose decline could be explained in no other way attested the tiny contribution made by autonomous changes in the virulence of disease organisms. This framework continues to shape discussion about the causes of mortality decline. It has, however, been criticized on both methodological and conceptual grounds. 2 Problems of measurement, aggregation, and interpretation have been identified. Can the independent operation of these major controls on mortality really be measured by the relative contributions that drops in mortality from tuberculosis, diarrhea, and smallpox made to decreases in the standardized mortality rates? Doesn't the national picture mask a set of important differences along, for example, rural/urban lines? Shouldn't historians be looking for interactions between the factors controlling mortality - interactions that might render invalid the isolation of factors and their unique identification with particular causes of death? Is it really acceptable to exclude all reference to behavioral factors, such as health care practices, from the analysis?