A study has been made of the bacterial flora in the sewage supplied to, and the effluent from, the bacteria beds which deal with the liquors produced at the Coventry gasworks. A high count in the effluent from the beds of bacteria containing an oxidase system, or a relatively large increase in the viability of such organisms when compared with the original sewage, has been found to be concurrent with good purification in the beds, and vice versa.

The bacterial flora undergoes considerable modification in the beds, and also when mixed with 4 per cent, of its volume of spent liquor in the laboratory. The total viable count of the treated sewage is reduced under these conditions, but the count of those bacteria which give the direct oxidase reaction and therefore catalyse the oxidation of catechol is relatively increased.

A gram negative vibrio (strain 0 1) is the only bacterium isolated which has proved capable of breaking down monohydric phenols. It has been constantly isolated from the effluent liquors.

Gasworks' ammonia liquor in the relatively high concentrations used exerts a markedly toxic action on all types of sewage bacteria, and the purification of this liquor by the beds is not so satisfactory. The toxicity is probably caused by the highly alkaline reaction of the liquor.

This investigation was commenced by one of us (F. C. H.) in conjunction with Dr Allan C. Monkhouse, who was at the time Research Chemist to the Liquor Effluents Sub-Committee of the Institution of Gas Engineers. Dr Monkhouse accepted another appointment shortly after the commencement of the work, and was therefore unable to assist further, but we should like to express our indebtedness to him during the initial period. We are also indebted to Prof. J. W. McLeod and Prof. J. W. Cobb for their kindly interest and encouragement, and to Mr P. N. Langford, Engineer and Manager of the Coventry Gas Department, for the facilities he so generously placed at our disposal at Coventry.

One of us (F. C. H.)is in receipt of grants from the Medical Research Council, and would again express his thanks to that body.

One of us (A. K.) is indebted to the Liquor Effluents and Ammonia Sub-Committee of the Institution of Gas Engineers for permission to publish this work.