The precursor-product relationship of very low density (VLDL) and low density lipoproteins (LDL) was studied. VLDL obtained from normal (NTG) and hypertriglyceridemic (HTG) subjects was fractionated by zonal ultracentrifugation and subjected to in vitro lipolysis. The individual subfractions and their isolated lipolysis products, as well as IDL and LDL, were rigorously characterized. A striking difference in the contribution of cholesteryl ester to VLDL is noted. In NTG subfractions, the cholesteryl ester to protein ratio increases with decreasing density (VLDL-I----VLDL-III). This is the expected result of triglyceride loss through lipolysis and cholesteryl ester gain through core-lipid transfer protein action. In HTG subfractions there is an abnormal enrichment of cholesteryl esters that is most marked in VLDL-I and nearly absent in VLDL-III. Thus, the trend of the cholesteryl ester to protein ratios is reversed, being highest in HTG-VLDL-I and lowest in VLDL-III. This is incompatible with the precursor-product relationship described by the VLDL----IDL----LDL cascade. In vitro lipolysis studies support the conclusion that not all HTG-VLDL can be metabolized to LDL. While all NTG subfractions yield products that are LDL-like in size, density, and composition, only HTG-VLDL-III, whose composition is most similar to normal, does so. HTG VLDL-I and VLDL-II products are large and light populations that are highly enriched in cholesteryl ester. We suggest that this abnormal enrichment of HTG-VLDL with cholesteryl ester results from the prolonged action of core-lipid transfer protein on the slowly metabolized VLDL mass. This excess cholesteryl ester load, unaffected by the process of VLDL catabolism, remains entrapped within the abnormal particle. Therefore, lipolysis yields an abnormal, cholesteryl ester-rich product that can never become LDL.