1. [75Se]selenite of [75Se]selenomethionine was injected in to the coelomic cavity of fish. After 2 d or 14 d the muscle portion of the fish was removed and homogenized. The long-term fate in rats of an oral dose of each labelled homogenate was compared with that of an oral dose of [75Ee]selenite or [75Se]selenomethionine mixed with unlabelled fish homogenate. 2. Urinary and faecal radioactivity were measured during the 1st week and whole-body radioactivity was determined for 10 weeks. Rats were killed at weekly intervals for 4 weeks for analysis of tissue distribution of 75Se. 3. Intestinal absorption of 75Se given as labelled fish homogenate was less complete than that of 75Se mixed with unlabelled homogenate, and the absorption of 75Se from the 14 d-labelled fish homogenate derived from [75Se]selenite was less complete than that of 75Se from the other labelled homogenates. 4. Urinary excretion of absorbed 75Se in the first 7 d was in the range 5--8% absorbed dose and was slightly greater in the rats given 75Se as selenite or derived from selenite than in those given 75Se as selenomethionine or derived from selenomethionine. Endogenous faecal excretion of absorbed Se was similar in all groups, as also were tissue distribution of retained 75Se and long-term whole-body turnover rate. 5. The results of these studies are compared with those of earlier studies of the metabolism in rats of [75Se]selenomethionine, [75Se]selenite, [75Se]selenocystine and 75Se incorporated in vivo into rabbit kidney. There were differences in the initial utilization of 75Se from these various sources but after the 1st week 75Se from all sources appeared to be metabolized similarly, suggesting that for rats dietary Se of all forms is ultimately incorporated into the same metabolic pool.