The bacteriological examinations of 1616 samples of polluted water from the Melbourne supply revealed that the production of acid and gas in MacConkey or in lactose broth was not always evidence of the presence of a coliform bacillus. There was, however, a marked difference in the number of confirmed samples in the two media. Among 1136 samples incubated in MacConkey broth coliform organisms were isolated by plating in 98·4%, but among 480 samples enriched in lactose broth 10·8% failed to confirm. MacConkey's broth is therefore recommended as the better medium for the presumptive coliform test in Melbourne water.The technique used in the examination of 1345 samples of water and 73 samples of animal faeces was Wilson's method IV extended by plating from at least one tube giving a positive presumptive reaction at 37° C. to include method I also. In both series of tests coli I appeared to be by far the more numerous type by method I, but when method IV was used the intermediate, aerogenes, cloacae group (abbreviated to I.A.C.) was detected in almost as many samples in water and in 63% of the faeces where the numbers were often very low compared with coli I. Unless I.A.C. was present alone or in large numbers in relation to coli I it was frequently missed by the plating method. Even coli I was more often found by method IV which proved, therefore, to be a more delicate test than method I for both types of organism.The 44° C. test for coli I proved to be highly specific in Victoria, since few organisms were able to produce acid and gas in MacConkey broth at 44° C. except coli I and organisms which had affinity with coli I rather than I.A.C. This contrasts with results reported from Madras (Raghavachari, 1939) and Singapore (Boizot, 1941) where the test appears to be less specific, but is in agreement with British (Bardsley, 1938; Batty-Smith, 1942) and Argentine (Ferramola, 1940) experience where the 44° C. test has also proved to be of high differential value.Growth in citrate in method IV was sometimes due to coli I, non-lactose fermenters or Pseudomonas pyocyanea, particularly when lactose broth was the enrichment medium.Bact. coli II and irregular I were found in heavily and in lightly polluted waters, and were associated rather more often with coli I than with I.A.C. They were seldom found in animal faeces. Irregular II rarely occurred among the organisms cultured, but the evidence suggests that it may be of faecal origin.The coliform counts on decaying leaves, algae and other vegetable debris found in the sumps and on the screens were very high, and the figures showed a heavy increase after incubation.Periodic sampling of cow dung left to dry in the open meadow showed that reduction in coliform count was very gradual and largely dependent on the moisture content, and that coli I was still overwhelmingly dominant 3–7 weeks after the samples were voided. Laboratory experiments, in which cow faeces were stored in moist and in dry chambers, gave similar results.Soil plots low in organic matter were watered with cultures of coli I, intermediate I or aerogenes I and sampled at intervals. All three types tended to die out gradually and there was no question of intermediate I or aerogenes I growing and multiplying under the conditions of the experiment, neither were any profound changes apparent in the biochemical reactions of the three types as a result of prolonged sojourn in soil. Soil which had been heavily contaminated with cow manure 3 years previously still contained large numbers of viable coli I.