Abstract Differential spectroscopic studies of complexes between glutamate dehydrogenase and its various substrates, reaction products, cofactors, regulatory nucleotides, and analogues of these compounds have permitted direct observation of the interactions of individual chromophoric groups of a given ligand with other ligands on the enzyme surface and with that surface itself. This ability to resolve multiple modes of binding into their individual components and comparison of the concentration dependence of these specific interactions with reaction kinetics lead to a model capable of explaining the complex effects of ligands on the reaction. The model represents an active patch of enzyme surface no larger than 12 by 25 A containing an array of six subsites each capable of binding a more or less specific functional group. These subsites and their respective specifications are Subsite I, which binds an intact amide group of nicotinamide as well as an adenosine group; Subsite II, which is specific for some portion of the pyrophosphate-ribose moiety of the adenylate group; Subsite III, which binds only a 5'-substituted pyrophosphate group of adenosine analogues; Subsite IV, which binds ammonium ion and the -NH3+ group of amino acids; Subsites Vα and Vγ, which are specific for carboxyl groups; Subsite VI, which binds the pyrophosphate group of GTP. Each ligand binding site is formed from some combination of these subsites; the complexity of the interactions results from the ability of some ligands to bind to more than one combination of subsites and from the fact that some subsites are common to two different ligands. TPNH binds to Subsites I and II, and the bound coenzyme itself contains all or some part of Subsites IV and Vα, explaining the obligatory order of the reaction and the cooperative binding of TPNH and l-glutamate. ADP and DPNH bind to Subsites I and III, accounting for the activation of the reaction by those ligands. GTP binds to Subsite VI and competes with ammonium for Subsite IV. While ADP and GTP share no single common subsite, their ligand sites cross in such a way that the two ligands mutually exclude each other by steric hindrance, demonstrating an allosteric effect not mediated by a conformational change. The model also provides a detailed and straightforward explanation of the differences in the kinetics of the reaction with TPN, DPN, 3-acetylpyridine DPN, and NMN.