1. A 25% faecal suspension in sodium chloride solution, incubated anaerobically at 37°C for 48 h, showed excellent survival of all the main groups of faecal bacteria. 2. All faecal incubation systems studied generated large amounts of ammonia, particularly those in which bacterial counts fell during incubation. As normal faeces contain negligible amounts of urea this ammonia must have been generated from sources other than urea. 3. Ammonia was also generated by faeces delivered by sodium chloride enema, and by ileostomy fluid, indicating that the phenomenon is not confined to distal colonic contents. 4. Ammonia generation by incubated faeces was inhibited by prior autoclaving of the sample, but not by sterilization with gamma-irradiation. 5. Generation of ammonia by incubated stool was accompanied by release of large amounts of organic anion and a fall in pH. 6. These observations are interpreted as evidence that ammonia generated within the colon in situ is not derived exclusively from urea, but also from bacterial deamination of amino acids, peptides and proteins. Simultaneously bacterial activity generates large amounts of organic acid. The presence of living bacteria is not essential for ammonia generation, provided that bacterial enzymes are present. 7. Bacterial generation of organic solute in faeces which have left the body is sufficiently rapid to cast serious doubts on the validity of faecal centrifugation, or other time-consuming techniques involving lengthy handling of faeces, as methods of obtaining extracellular faecal fluid for measurements of organic constituents or ammonia.