This investigation has the overall aim of elucidating the basic mechanism of membrane fusion using model systems. It involves two related but independent investigations. In the first part of the investigation, it was found that protein 4.1 from human erythrocyte membrane induces the fusion of phophatidylserine (PS)/phosphatidylethanolamine(PE) vesicles under acidic condition and causes the leakage of the vesicles at a broad pH range. The pH-dependent binding of protein 4.1 to phospholipid vesicles was similar, but not identical to the pH-dependent fusion. Proteolytic digestion of vesicle-protein 4.1 complex with trypsin showed that segments of protein 4.1 with molecular weight of approximately 26,000 and 33,000, respectively, are protected from the proteolysis. Protection of the vesicles from the attack of phospholipase D in the presence of protein 4.1 at neutral pH showed that protein 4.1 makes a close contact with the surface of vesicles. It does not, however, bind irreversibly to the vesicles at the pH. In the second phase of this investigation, fusion of phosphlipid vesicles compoesd of various combination of phosphatidylcholine, phosphatidylethanolamine and phosphatidic acid was induced by $Ca^{2+}$ with of without the presence of phospholipase D at neutral pH. For the vesicles composed of 20:50:30 molar ratio of phosphatidylcholine/phosphatidylethanolamine/phosphatidic acid, the initial fusion rate was much faster than the expected value when only the conversion of phosphatidycholine and phosphatidylethanolamine into phosphatidic acid by phospholipase D is taken into account. Inhibition of phospholipase D activity also drastically reduced the extent of fusion to the level of $Ca^{2+}$-induced fusion. These observations are discussed in terms of the involment of outer monolayer of the vesicles and the enzyme activity itself in the fusion process.