Mathematical models are developed to aid in the investigation of the implications of heterogeneity in contact with infection within a community, on the design of mass vaccination programmes for the control of childhood viral and bacterial infections in developed countries. Analyses are focused on age-dependency in the rate at which individuals acquire infection, the question of 'who acquires infection from whom', and the implications of genetic variability in susceptibility to infection. Throughout, theoretical predictions are based on parameter estimates obtained from epidemiological studies and are compared with observed temporal trends in disease incidence and age-stratified serological profiles. Analysis of case notification records and serological data suggest that the rate at which individuals acquire many common infections changes from medium to high and then to low levels in the infant, child and teenage plus adult age groups respectively. Such apparent age-dependency in attack rate acts to reduce slightly the predicted levels of herd immunity required for the eradication of infections such as measles, when compared with the predictions of models based on age-independent transmission. The action of maternally derived immunity in prohibiting vaccination in infants, and the broad span of age classes over which vaccination currently takes place in the U.K., however, argue that levels of herd immunity of between 90 and 94% would be required to eliminate measles. Problems surrounding the interpretation of apparent age-related trends in the acquisition of infection and their relevance to the design of vaccination programmes, are discussed in relation to the possible role of genetically based variation in susceptibility to infection and observations on epidemics in 'virgin' populations. Heterogeneous mixing models provide predictions of changes in serology and disease incidence under the impact of mass vaccination which well mirror observed trends in England and Wales.