Over a period of about 10 months samples were taken at frequent intervals from various stages of treatment at two sewage works; one of these was a small rural works treating only domestic sewage, and the other treated the sewage of an industrial town. These samples were submitted to bacteriological and to chemical examination.

The total plate counts, the counts of coli-aerogenes bacteria, and the counts of Streptococcus faecalis in settled sewage were higher in the warmer than in the colder months of the year.

Though the average total numbers of bacteria in the sewage containing trade wastes were appreciably higher than that of the domestic sewage the counts of faecal organisms were somewhat lower.

Treatment at the works considerably reduced not only the polluting strength of the sewage but also the numbers of bacteria. At the works receiving sewage from an industrial town an average of about 80% of bacteria in settled sewage were removed by treatment in the bio-aeration plant, percolating filters, and humus tanks, and a further 8–10% were removed by subsequent treatment in pressure sand or anthracite filters. At the small rural works, where the rate of treatment was much lower, rather more than 90% of the bacteria were, on the average, removed by treatment in percolating filters and humus tanks, and a further 6% by subsequent passage through shallow gravity sand filters.

The final effluents discharged from the works still contained large numbers of bacteria. Their effect on the bacterial content of the streams to which they are discharged may be gauged from the fact that, in the effluent from the small rural sewage works, which was of consistently good quality when judged by chemical criteria, there were on the average about 160,000 cells of faecal Bacterium coli and about 20,000 faecal streptococci per 100 ml. Numbers in the effluent from the works in the industrial town were even higher. Sand filters of the type investigated here could therefore not be relied upon to produce an effluent of good bacterial quality.

The 44°C. test in MacConkey broth, prescribed by the Ministry of Health (1939) for examination of water, was found, when applied to samples of sewage or sewage effluent, to be almost specific for Bacterium coli, type I. Growth with formation of acid in sodium azide broth incubated at 45° C. was found to be an almost specific test for Streptococcus faecalis.