To study membrane biogenesis and to test the validity of the endomembrane flow hypothesis, incorporation of 32P and [Me-3H]choline in vivo into membranes of the rat liver was followed. Rough microsomal, Golgi-rich, and plasma membrane fractions were monitored with marker enzyme assays and shown with morphometric analysis to contain 82% rough microsomes, at least 70% Golgi complexes, and 88% plasma membranes, respectively. Membrane subfractions from the rough microsomal and Golgi-rich fractions were prepared by sonic disruption.At 5 to 30 min after 32P injection, the specific radioactivity of phosphatidylcholine was higher in the rough microsomal membranes than in the Golgi membranes. From 1 to 3 h, the specific activity of phosphatidylcholine in Golgi membranes became higher and reached the maximum at about 3 h. Although the plasma membrane had the lowest specific radioactivity throughout 0.25–3 h, it increased rapidly thereafter to attain the highest specific activity at 5 h. Both rough microsomal and plasma membranes reached their maxima at 5 h.The specific radioactivity of [32P]phosphatidylethanolamine in the three membrane fractions was similar to that of [32P]phosphatidylcholine except from 5 to 30 min, when the specific radioactivity of phosphatidylethanolamine in the Golgi membranes was similar to the rough microsomal membranes.At 15 min to 5 h after [Me-3H]choline injection, more than 90% of the radioactivity in all the membranes was acid-precipitable. The specific radioactivities of the acid-precipitated membranes, expressed as dpm per milligram protein, reached the maximum at 3 h. After [Me-3H]choline injection, the specific radioactivity of phosphatidylcholine separated from the lipid extract of the acid-precipitated membranes (dpm per micromole phosphorus) did not differ significantly in the three membrane fractions. The results indicated rapid incorporation of choline into membrane phosphatidylcholine by the rough endoplasmic reticulum, Golgi, and plasma membranes simultaneously.The data with both 32P and [Me-3H]choline precursors did not support the endomembrane flow hypothesis. The Golgi complexes apparently synthesized phosphatidylethanolamine and incorporated choline into phosphatidylcholine as well as the endoplasmic reticulum. The results are discussed with relevance to current hypotheses on the biogenesis and transfer of membrane phospholipids.