One of the many factors which have been implicated in the aetiology of SIDS is infection of the respiratory tract, particularly viral infection. This applies particularly to those infants who die from SIDS who are more than 3 months old. The evidence for this belief is based on both epidemiological and pathological factors. Among the epidemiological factors are the pronounced seasonal variation of SIDS (it being commoner in winter); the increased incidence of pre-existing illness, particularly upper respiratory infections, in the two weeks before death; and the increased occurrence of SIDS during epidemics of viral infection in the community. Not all of these factors are universally accepted, however, particularly when appropriate controls are investigated. The necropsy evidence includes the presence of lymphoid inflammatory infiltrates in the respiratory tract, particularly the upper respiratory tract. While these are present in many cases of SIDS, they are not present in all. Postmortem isolation of respiratory viruses has also given conflicting results: some authors show an apparent increase compared with controls, while others do not. No specific virus has been implicated. Part of the reason for these conflicting epidemiological and pathological results is failure to use proper controls. An additional explanation may be the technical difficulties involved in isolating viruses. Apart from the problems resulting from postmortem effects, culture, immunofluorescence, and ELISA tests are known to give significant false negative rates. Accordingly, newer, potentially more sensitive and robust techniques, such as molecular hybridisation, are being applied to cases of SIDS to determine whether viral infection is more common than is currently recognised. Whatever the outcome of these investigations, it is highly unlikely that viral infection per se is the cause of SIDS. One or more additional factors are also involved which may include an abnormal immune response, generation of thermal stress, precipitation of respiratory obstruction, bacterial overgrowth with toxin release, or suppression of the arousal response.