Fresh undiluted sewage contains unstable substances which are directly toxic to fish. Sewage which has been kept for 1 or 2 days and stirred in an open tank at 18° C. is more toxic than fresh sewage, possibly owing to the production of other toxic substances. It has also been shown that the biochemical oxygen demand is directly proportional to the "strength" of the sewage which is closely related to toxicity. Consequently a large biochemical oxygen demand of fresh sewage indicates a certain degree of toxicity.
It appears that when fresh sewage is discharged into a river in concentrations up to 10 per cent. of the river water, it has no directly toxic effect on fish. A river can purify a certain proportion of sewage effluent without becoming seriously deoxygenated. In one river with which we are acquainted a sewage effluent forms about 3.7 per cent. of the river flow at normal summer levels. The dissolved oxygen content of the river during the period it was under observation never fell below 60 per cent. of the saturation value, and such a mixture of sewage and water would have no directly injurious effect on fish. If, however, the proportion of sewage is increased, then the river will become more deoxygenated and eventually at a certain ratio of sewage to river water the toxicity will become more apparent, partly owing to the oxygen depletion. This condition is of course readily produced in a small river. The final stage will be attained when the quantity of sewage is sufficient to remove all the dissolved oxygen of the water. When there is a belt of deoxygenated water, sewage will decompose under anaerobic conditions and sulphide will be formed. As it has been shown that this substance is directly toxic to trout in small concentrations, 0.3 part per 100,000 being rapidly fatal, the fish are exposed to a definite poison as well as the danger of asphyxiation. In a slow-flowing river below a sewage works, mud is deposited on the river bed. Anaerobic conditions can exist in this mud although the supernatant water is fully oxygenated and any sulphate present may be reduced to sulphide which will be washed downstream when the river is in flood.
It may therefore be concluded that in a river where there is a plentiful supply of oxygen, sewage in moderate quantity has no directly harmful effect on fish unless it is in a septic condition and contains sulphide. When sewage is discharged into a slowflowing river, mud banks will be formed which may have the necessary anaerobic conditions for the production of sulphide. These, during floods, may be harmful to fish before they are oxidised. There remains the possibility that, although the fish are not directly affected, their food and spawning grounds may suffer from the pollution. This possibility has not been overlooked, but the data on these points are not yet complete.