Information on salmonella serovar, community, and date of isolation concerning 10368 cases of S. enteritidis (SE) infection and 1163 cases of 5. typhimurium (ST) infection reported in 1992 to the Salmonella Data Bank of the Land of Brandenburg have been analyzed for clusters and trends on case/time diagrams. A breakdown by district provided a suitable degree of ''resolution'' for discrimination of ''Episodes of Increased Case Frequency'' (EICF) and 7 days moving averages. Thus 39 of the 44 districts of the Land were included in the evaluation of SE cases. The population size of the remaining 5 districts was too high for clear episode discrimination. EICFs were subdivided into (a) 16 outbreaks (another 10 outbreaks occurred outside EICFs), which had also been identified and officially reported by local health services as a result of compulsory investigation of cases and their social environment, and (b) 44 ''case accumulations'' (ACCs), which were only identified by retrospective analysis of the case/time diagrams. By this procedure, three case categories of distinct properties have been found for SE transmission from the animal reservoir to the human population: outbreaks, ACCs, and ''truly sporadic cases'' scattered over the district outside outbreaks and ACCs (Table 3). Truly sporadic cases (Table 6) represented the largest proportion of SE cases. ACCs accounted for 37% and outbreaks, for 12%. The most striking result concerns the ACCs, which cover slightly larger geographical areas (mean: 3.5 communities) and last longer (mean: 3 weeks, but up to 7 weeks) than officially recorded outbreaks (means: 2 communities and 2 weeks). SE-ACCs differed from SE outbreaks also in the seasonal pattern Otherwise, ACCs met outbreak characteristics including the average number of persons involved (mean: 64, for outbreaks: 42; Table 4). The results suggest that SE-ACCs represent an important and distinct element in the SE epidemic. This conclusion is supported by a comparison of the findings with results obtained by the same procedure for ST cases. ST outbreaks do not differ from ST-ACCs in any respect, and both types of ST-episodes are of significantly shorter duration and comprise less people than episodes due to SE. The public health significance of the SE-ACCs as a newly detected and particular element of the current SE epidemic is discussed. In view of the epidemiological significance, i.e. the high proportion of all SE cases involved, it appears justified to develop methods for the early detection of ACCs and the search for the ''missing link'' between the cases and their source at the community level.