The existence of chemical changes in the brain correlated with acquisition and storage of information, and impairment of learning by inhibitors of RNA and protein synthesis suggest that chemical processes are involved in long‐term memory. The main controversy is over the significance of these processes: whether they are merely concerned with increased metabolic needs or represent a molecular code in which neural information would be processed and preserved. The molecular coding hypothesis can neither be proved nor disproved by a purely chemical approach; additional means are necessary for revealing the presence of minute amounts of specific substances. One of these means is the behavioral bioassay which, in the last ten years, has guided the detection of eight learning‐induced substances, all peptides, in the brain. Two of them (scotophobin, ameletin) have been fully identified and synthesized; they can now be determined and studied by chemical methods. Three other peptides have been purified and their structure is being elucidated. Although good correlation has been found between scotophobin and dark avoidance, specificity of each peptide for a given learned behavior requires further study. The nature of the hypothetical coding system is unknown but it operates probably within the framework of neural circuits: it could create new synaptic junctions, mark them and guide the flow of nerve impulses along the new channels induced by learning. This reprogramming may be based on the chemical recognition system that is believed to direct the formation of the genetically programmed pathways.
