SUMMARY The effect of lactulose on faecal pH and ammonia has been studied in three normal subjects with the aid of dialysis of faeces in vivo. Observations were also made with sodium sulphate and the two hexahydric alcohols, mannitol and sorbitol, given in doses sufficient to cause a similar increase in stool weight.All four cathartics rendered the stool more acid, but there was no increase in the concentration of faecal ammonia. Lactulose, despite increasing faecal volume, did not cause an increase in the absolute amount of ammonia lost in the faeces, but the other purgatives did show a modest rise.The results are inconsistent with the theory that lactulose benefits the clinical picture of portosystemic encephalopathy by trapping ammonia in an acid stool. An alternative suggestion is advanced, namely, that any cathartic (including lactulose) reduces ammonia absorption from the colon by decreasing colonic transit time, and so reducing the amount of ammonia generated by autolysis of colonic bacteria.Lactulose (1-4 ,B galactosido-fructose) was introduced in 1966 by Bircher, Muller, Guggenheim, and Haemmerli, in the treatment of portosystemic encephalopathy. This step had been suggested by Ingelfinger (1964-65) who recognized that the disaccharide is not hydrolysed in the ileum and when taken by mouth will reach the colon unchanged, where it promotes the growth of lactobacilli, organisms which lack urease, and other ammonia-generating enzymes (Phear and Ruebner, 1956;Macbeth, Kass, and McDermott, 1965). Ingelfinger (1964) suspected that a reduction in the number of proteolytic bacteria in the colon, in favour of lactobacilli, would reduce the formation of ammonia, leading to a decrease in portal venous ammonia.Clinical trials have since tended to confirm the value of lactulose in portosystemic encephalopathy (Bircher, Haemmerli, Scollo-Lavizzari, and Hoffmann, 1971a). Lactulose has also been shown to lower faecal pH (Mayerhofer and Petuely, 1959), a finding which has led to an alternative suggestion, that its effect on non-ionic diffusion of ammonia might explain its efficacy in portosystemic encephalopathy. It was argued (Elkington, Floch, and Conn,