The time-averaged lateral organization of the lipids and proteins that make up mammalian cell membranes continues to be the subject of intense interest and debate. Since the introduction of the fluid mosaic model almost 50 years ago, the "lipid raft hypothesis" has emerged as a popular concept that has captured the imagination of a large segment of the biomembrane community. In particular, the notion that lipid rafts play a pivotal role in cellular processes such as signal transduction and membrane protein trafficking is now favored by many investigators. Despite the attractiveness of lipid rafts, their composition, size, lifetime, biological function, and even the very existence remain controversial. The central tenet that underlies this hypothesis is that cholesterol and high-melting lipids have favorable interactions (i.e., they pull together), which lead to transient domains. Recent nearest-neighbor recognition (NNR) studies have expanded the lipid raft hypothesis to include the influence that low-melting lipids have on the organization of lipid membranes. Specifically, it has been found that mimics of cholesterol and high-melting lipids are repelled (i.e., pushed away) by low-melting lipids in fluid bilayers. The picture that has emerged from our NNR studies is that lipid mixing is governed by a balance of these "push and pull" forces, which maximizes the number of hydrocarbon contacts and attractive van der Waals interactions within the membrane. The power of the NNR methodology is that it allows one to probe these push/pull interaction energies that are measured in tens of calories per mole.