Membrane fusion is an essential step in the infection of permissive cells with human immunodeficiency virus (HIV). Infected cells frequently fuse with each other, and then progress to form multinucleated giant cells (syncytia). To gain insight into mechanisms of HIV env-mediated membrane fusion, we developed a new assay for studying the initial events. The assay is based on the redistribution of fluorescent markers between membranes and cytoplasm of adjacent cells examined by means of fluorescence video microscopy. Membrane fusion between HIV-1 envelope glycoprotein (gp120/41) expressing effector cells and CD4+ target cells was observed 90 min after the association of cells, whereas the first syncytia only became apparent after 5 h. Moreover, membrane fusion events were observed under conditions where no syncytia were detected, for example, when the effector:target cell ratio was greater than 100:1, or less than 1:100. A significant number of cells with fused membranes were not involved in the syncytia. In order to determine whether quantitative differences in receptor expression might influence the extent of membrane fusion, we used laboratory-selected variants of CEM cells that differ in their expression of CD4. We found that CD4 is required on the target membrane for HIV env-mediated membrane fusion, but its extent is only partially dependent on CD4 surface concentration. The ability of those CEM variants to take part in HIV env-mediated membrane fusion did not correlate with their capacity to form syncytia. These findings indicate that additional steps are needed to form syncytia after membrane fusion.