Our earlier work demonstrated that the rate of protein synthesis in the exocrine cells of the rat pancreas is constant in different physiological states, including prolonged fasting. In this study we have followed the fate of the protein in the pancreatic cells of the fasting animal in vivo as well as in vitro. The data were obtained by quantitative radioautography and by biochemical determinations. In nonanesthesized, fasting rats, without cannulated pancreatic duct, some 80% of the proteins synthesized at a given time leaves the cell within 12 hr by way of secretion, intracellular breakdown not being important. Two mechanisms of fasting secretion exist. The first, starting at a slow rate after 20 min, is inferred to result from fortuitous contacts of young secretory granules with the apical cell membrane. The rate of secretion is the same in vivo as in vitro, at least during the first 4 hr after pulse labeling. Within 7 hr about 20% of the total amount of newly synthesized protein has left the cell. The second mechanism consists of an orderly movement of the mass of secretory granules towards the apical cell membrane as caused by the continuous assembly of new granules. The granules that come into contact with the cell membrane are discharged. It takes about 7-12 hr for secretory protein transported in this way to reach the cell membrane. The addition of new secretory granules to those present is essential for the second mechanism, for the blockade of protein synthesis by cycloheximide decreases the rate of this phase of secretion without interfering with the secretory process proper. Atropin does not inhibit the fasting secretion in vitro, nor does extensive washing of the tissue slices, excluding possible secretagogues as important factors in fasting secretion.