Vaccination is but one element in a control programme for bovine respiratory disease. Its laboratory study can be divorced from the others but its field application cannot. The problems associated with the development of effective vaccines fall into two broad groups: multiplicity and ubiquity of pathogens and secondly the identification of the crucial elements in an immune response. Agricultural systems which experience annual outbreaks of respiratory disease attributable to the same pathogen in cattle of specific age have the choice of using passive or active immunity of minimal valency. In the majority of systems the cause and timing of an outbreak cannot be predicted and therefore multivalent vaccines are required. Both inactivated and modified live products are available for use against the well-known pathogens. Their relative advantages hinge on the significance attributed to the ability to stimulate the production of particular immunoglobulins at specific body sites and the persistence of the responses. The widely held view that success requires the stimulation of secretory antibodies by intranasal administration of living vaccines is not universally accepted. An assessment of their protective value is not easily made because of the difficulty of reproducing an adequate field challenge in the laboratory. The measurement of serological responses and virus shedding times following challenge are of limited value as alternatives.