Recovery rates of fecal coliforms and of Escherichia coli were determined at 44.5, 45.0 and 45.5°C in raw milk, ground meat and raw sewage. MPN values based on gas production (Standard MPNs) and on gas production and/or growth only (Total MPNs) were calculated for both fecal coliforms and for E. coli . The expected trend towards lower MPN values with increasing incubation temperature was more pronounced for the Standard MPNs than for the Total MPNs. The temperature effect was also strongly product - specific in that the Total and Standard MPNs for the fecal coliforms and the Standard MPNs for E. coli for sewage only differed significantly from one another within each determination at the three different incubation temperatures. The effect of length of incubation time on the ratios of E. coli to fecal coliforms was most pronounced at 45.5°C. Product specificity was again observed. The greatest increase in the recovery rate of aerogenic E. coli between 24 and 48 h of incubation time occurred in sewage (66%). For meat, the increase was 57% and for milk 46%. In terms of combined (aerogenic and anaerogenic) E. coli (expressed as Total MPNs), the increases were considerably less, but highest for the meat (33%), followed by sewage (29%) and by milk (21%). A breakdown of the E. coli isolates recovered from both gas-positive and gas-negative primary (fecal coliform) EC broth tubes showed that for the three products combined there were eight times as many false-positives at 44.5°C as at the other two incubation temperatures. In contrast, there were 12% false-negatives at 45.5°C compared to 3% at 45.0°C and 2% at 44.5°C. Since the high incidence of false-negatives (loss of E. coil ) at 45.5°C is not counter-balanced by an enhanced specificity (fewer false-positives) over 45.0°C, the latter temperature is to be preferred. Meat yielded the lowest rate for false-positives at any of the three incubation temperatures. In contrast, at 45.5°C, it gave 21% false-negatives compared to only 9% for sewage and 10% for milk. On the other hand, milk contributed the most false-positives at 44.5°C (20%), compared to only 1% for meat and 3% for sewage. A potential loss of 21% of E. coli - containing EC broth tubes is hardly tolerable, reinforcing the contention that gas formation at evaluated temperatures is not a valid criterion of fecal origin.