||Bacteria swim through liquid with flagella; each flagellum is a rigid helix spun by a rotary motor at its base, powered by the transmembrane proton gradient (or Na+ gradient in some species). The motor is significant as a virulence factor, as a complex biochemical system with important evolutionary connections, and as a self-assembling molecular machine to inform nanoengineering efforts. Thoroughly understanding motor function demands detailed structural information. In Escherichia coli and Salmonella enterica serovar Typhimurium, the flagellar motor is ~45 nm in diameter and contains ~25 different proteins, some present in multiple copies. The rotating inner portion of the motor (the rotor) contains many copies of the proteins FliM, FliG, and FliN, all of which are essential to rotation and direction switching. Partial X-ray crystal structures are available for these rotor proteins, but spatial relationships between the proteins must also be understood. Electron microscopy provides structural information at a larger scale, but the resolution is not yet sufficient to dock crystal structures. To address the need for intermolecular structural data, I used the biochemical technique of site-directed cross-linking (SDCL) to find the rough relative placement between FliG and FliG, between FliM and FliM, and between FliM and FliG. Guided by crystal structures, I used site-directed mutagenesis to replace one or two surface residues with chemically reactive cysteine (Cys). Oxidizing or sulfhydryl-specific agents then induced cross-linking between those Cys residues close enough to react together. Assays of swimming ability controlled for disrupted motor assembly or function. Results indicate that adjacent FliG middle domains are close together with nothing intervening, and place constraints on the relative orientations of the domains. A study of FliM middle domains led to a similar conclusion about spacing and provide constraints for these portions. Results for FliG C-terminal domains imply larger relative spacing with significant conformation freedom. Preliminary results for FliG-FliM cross-linking are presented. Data on FliG-FliM binding are reviewed, along with a recent model (proposed by others) that FliG-FliM binding alternates depending on position. I propose a functional model based on this binding model and discuss considerations of the functional model.