Strains of Listeria monocytogenes from 475 cases of human listeriosis collected during 1967-1984, belonged to one of three serogroups (1/2, 3 or 4). They were phage typed with a set of 28 phages to investigate three aspects of the epidemiology of listeriosis. Three patients each had two episodes of listeriosis, 3 months to 2 years apart, with strains of the same serogroup and indistinguishable by phage typing. Ten episodes of possible cross-infection between pairs of neonates in the same hospital occurred; the first baby was ill at or within 1 day of birth, and the second baby became ill 8-12 days after contact with the first. In each pair the L. monocytogenes strains were of the same serogroup and indistinguishable by phage typing. In three clusters of cases there may have been a common source of infection. L. monocytogenes strains from 10 of 11 cases of listeriosis in the Carlisle area in Jul.-Dec. 1981 were of the same serogroup; nine strains were non-phage-typable. The second cluster involved four adults treated at one hospital and the third a pair of neonates who were ill shortly after birth. In each cluster, strains were of the same serogroup, and were indistinguishable by phage typing. These last two clusters occurred during a short period when an unusually high proportion of strains from all cases of human listeriosis in Britain were indistinguishable by phage typing from the cluster strains, suggesting the possibility of common source infection.