The clinical syndrome of fat embolism can be a complication in patients who have suffered long-bone fracture, and has been associated with a number of conditions in which bone fracture is not a factor. The origin of embolic fat (bone marrow or depot fat versus plasma lipoproteins) has been debated. To test the hypothesis that alterations in the composition of plasma very-low-density lipoproteins (VLDL) could lead to their coalescence and subsequent impaction in pulmonary vasculature, New Zealand white rabbits were injected intramuscularly with 25 mg of cortisone twice a day for 2, 4, 6, 11 and 12 days. Blood samples were obtained from unanesthetized animals prior to the cortisone treatment and at sacrifice. Pulmonary fat embolism that was similar histologically to that found in the human patient with the clinical syndrome was produced in all rabbits treated for 4-12 days. After the fourth day, all animals were hyperlipidemic. There was a tenfold increase in plasma VLDL concentration and the mean particle diameter of the VLDL was increased from 620 A to 1016 A. The proportion of triglyceride (TG) was increased, protein (PR) and phospholipid (PL) were decreased and total cholesterol (TC) was unaltered (untreated controls: PR, 16.4%; TG, 58.5%; PL, 19.2%; TC, 6.7% and cortisone-treated: PR, 8.4%; TG, 75.1%; PL, 11%; TC, 6.4%). The lipid composition of pulmonary embolic fat from rabbits treated with cortisone for 6 or 11 days was TG, 81%; PL, 13% and TC, 6%. VLDL from similarly treated rabbits had a lipid composition of TG, 79%; PL, 14.5% and TC, 6.6%. The TC concentration of bone marrow and depot fat from animals treated with cortisone for 11 days was <1%, and the PL concentration was <2%. These data support the hypothesis that plasma lipoproteins, primarily the VLDL, when altered in size and composition can coalesce into globules of sufficient size to occlude peripheral capillaries.