Radioactivities were measured in serum, urine, and bile of dogs at different times after intravenous injection of 14C-methyl-gamma-amanitin (14C-A) and 3H-O-methyl-dehydroxymethyl-alpha-amanitin (3H-A). For either substance, the relation between the specific plasma activity C and the time t could be best described with the function C = C1 X e- lambda 1 X t + C2 X e- lambda 2 X t. Therefore the linear open two-compartment system was selected as an adequate toxicokinetic model. Most important, the distribution volumes (in the steady state) were in the range of the extracellular space, and the total body clearances were in the range of the dog creatinine clearance. In accordance with former findings for 3H-A, 14C-A was not bound to plasma proteins. More than 80% of 14C-A was eliminated in the urine; less than 10% was found in the bile. From these data, two suggestions may be derived for the therapy of Amanita intoxication in man. First, detection in the urine of amatoxins 2 or 3 days after mushroom ingestion points to an ongoing amatoxin absorption or reabsorption from the intestine, and should lead to therapy with adsorbents and, in the absence of diarrhea, with laxatives. Second, hemoperfusion will remove significant amounts of amatoxins during the time of ongoing absorption or reabsorption and a few hours thereafter.