1. The distribution of coliform bacteria in lakes and streams has been studied. Weekly samples have been collected from different depths from the north and south basins of Windermere, from Thirlmere, and Bsthwaite Water, dnd from streams flowing into these lakes. Nearly 300 cultures have been isolated from positive tubes of MacConkey broth and their relationship to the coliform group has been studied and considered.

2. In relatively unpolluted streams the counts of coliform bacteria and the plate counts on sodium casemate agar tended to fluctuate in the same direction as river level, but in polluted streams the increased flow of water accompanying a higher river level reduced the numbers of bacteria per unit volume. Counts of coliform bacteria were, in general, much higher in summer than in winter, despite the lower rainfall in summer.

3. During the winter period of circulation of the water, samples of water from different depths in the lakes gave approximately the same count for coliform bacteria. When stratification of the water became established numbers of coliform bacteria in the hypolimnion (the lower stratum) dropped to very low figures and remained fairly constant; much higher counts were found in the epilimnion (the upper stratum).

Large fluctuations in numbers of coliform bacteria which occurred from week to week in the lakes were related to fluctuations in plate counts at only one of the seven sampling points, and were not related to previous rainfall. The count of coliform bacteria was greatest in summer and autumn, a phenomenon that was by no means entirely a result of the bacteria which were washed in being concentrated in the epilimnion. It is suggested that temperature may have affected viability and proliferation.

4. Of the total number of coliform organisms producing acid and gas from lactose at 37° C, approximately 70% were Bact. coli types I and II, and the remainder were members of the I.A.C. group. In the relatively pure waters of Thirlmere and its inflows Bact. coli, type I, made up 86 and 98% respectively of the total coliform cultures isolated, whereas in the relatively impure waters of Windermere and Esthwaite Water the percentages were 39 and 37 respectively. The differences were due to the greater proportions of Bact. aerogenes, type I, and of intermediate and irregular types in the impure waters. The significance of greater proportions of the I.A.C. group in polluted rather than in unpolluted water is discussed. The actual counts of coliform bacteria were very much lower in Thirlmere than in the other lakes examined.

5. Of the total cultures isolated from positive tubes of MacConkey broth 27% did not produce acid and gas in MacConkey broth at 37° C. Approximately one-half of these cultures, however, fermented lactose at 30° C, and gave varied results with the standard differential tests. The majority of these cultures could, by gradually increasing the incubation temperatures of successive inoculations, be “trained” to ferment lactose at 37° C but not at 40° C. They were more frequently isolated from the polluted lakes and inflows than from unpolluted waters. The remaining organisms isolated proved to be non-lactose-fermenting species belonging to the genera Proteus, Salmonella, Eberthella and Alkaligenes (Bergey).