The lateral organization of fluid cholesterol-dimyristoylphosphatidylcholine (DMPC) bilayers was studied by measuring the response of fluorescent membrane probes, dipyrenylphosphatidylcholines (diPyrxPCs) or merocyanine 540, to the variation of cholesterol concentration. Parallel absorbance and light-scattering measurements were also carried out. The excimer-to-monomer ratio of diPyrxPCs displayed abrupt deviations at particular cholesterol mole fractions (CMFs). The most notable of these occurred at CMFs of 0.15, 0.33, and 0.67. Deviations were also frequently observed at CMFs of 0.12, 0.20, 0.25, and 0.40. Merocyanine 540 reproducibly reported deviations at CMFs of 0.15 and 0.33 and frequently reported values close to 0.12, 0.20, and 0.25. In absorbance (turbidity) and light scattering versus CMF plots, well-defined kinks were observed at CMFs of 0.16, 0.33, 0.52, and 0.67. The occurrence of kinks or other deviations at those particular CMFs is most readily explained in terms of a superlattice model previously developed to explain the lateral distribution of pyrenylphospholipids in bilayers [Somerharju, et al. (1985) Biochemistry 24, 2773-2781; Virtanen, J. A., et al. (1988) J. Mol. Electron. 4, 233-236]. This model is based on the assumptions that (i) each cholesterol molecule replaces a single acyl chain in a hexagonal lattice, (ii) cholesterol molecules, because of their larger size, perturb the lattice, (iii) this perturbation is minimized when the cholesterol molecules are maximally separated from each other, and (iv) the maximal separation is achieved when the cholesterol molecules form a hexagonal or centered rectangular superlattice. All detected critical CMFs, except that at CMF 0.67, are predicted by the model, thus strongly supporting its validity. The critical CMF at 0.67 is a limiting case, which can be accounted for by assuming that cholesterol and phospholipid molecules form alternating rows, i.e., formation of a cholesterol superlattice with rectangular symmetry. As predicted by the superlattice model, composition-driven order-to-disorder transitions occur between the critical CMFs, as indicated by increased data scatter and sample fluctuations in those regions. Another important prediction of the superlattice model is that domains with different cholesterol superlattices should coexist at most cholesterol concentrations. Such domains do not have to be extensive to account for the critical events observed here; rather, they are expected to be dynamic entities of limited size. It is very likely that such microscopic domains with distinct cholesterol superlattices also coexist in biological membranes. This is expected to have remarkable effects on both the structure and functions of these membranes.