Salmonella enterica serovar Typhimurium can differentiate into hyperflagellated swarmer cells on agar of an appropriate consistency (0.5 to 0.8%), allowing efficient colonization of the growth surface. Flagella are essential for this form of motility. In order to identify genes involved in swarming, we carried out extensive transposon mutagenesis of serovar Typhimurium, screening for those that had functional flagella yet were unable to swarm. A majority of these mutants were defective in lipopolysaccharide (LPS) synthesis, a large number were defective in chemotaxis, and some had defects in putative two-component signaling components. While the latter two classes were defective in swarmer cell differentiation, representative LPS mutants were not and could be rescued for swarming by external addition of a biosurfactant. A mutation in waaG (LPS core modification) secreted copious amounts of slime and showed a precocious swarming phenotype. We suggest that the O antigen improves surface "wettability" required for swarm colony expansion, that the LPS core could play a role in slime generation, and that multiple two-component systems cooperate to promote swarmer cell differentiation. The failure to identify specific swarming signals such as amino acids, pH changes, oxygen, iron starvation, increased viscosity, flagellar rotation, or autoinducers leads us to consider a model in which the external slime is itself both the signal and the milieu for swarming motility. The model explains the cell density dependence of the swarming phenomenon.