A total of 574 samples, of seven different types, were examined for the presence of salmonellas. All the specimens were pre-enriched in buffered peptone water and enriched in Rappaport-Vassiliadis medium (RV medium). In one trial 0.1 ml of pre-enrichment culture of 497 samples (79 chicken carcasses, 228 specimens of minced meat, 100 pork sausages, 19 samples of dried powdered chicken meat, 11 specimens of faeces of healthy pigs and 60 samples of sewage polluted natural sea water) was seeded to 10 ml as well as to 100 ml of RV medium. With the first inoculum (ratio 0.1:10), 111 samples were found to contain salmonellas, while with the second inoculum (ratio 0.1:100), only 102 positive samples were detected. This difference is marginally significant (P less than 0.05). In another trial, 0.1 ml, 0.2 ml and 0.5 ml of pre-enrichment culture of 162 specimens (71 chicken carcasses, 40 samples of sewage polluted sea water and 51 samples of sewage polluted river water) were in turn introduced to 10 ml of RV medium. With the 0.1 ml inoculum 93 positive samples were detected, while with the 0.2 and 0.5 ml inocula 93 and 88 positive samples were found. The differences are not statistically significant. In these trials the growth of competing organisms was minimal with ratios of inocula 0.1:10 or 0.1:100.