This article reviews especially the early history of glutamate and GABA as neurotransmitters in vertebrates. The proposal that some amino acids could mediate synaptic transmission in the CNS initially met with much resistance. Both GABA and its parent glutamate are abundant in the brain; but, unlike glutamate, GABA had no obvious metabolic function. By the late 1950s, the switch of interest from electrical to chemical transmission invigorated the search for central transmitters. Its identification with Factor I, a brain extract that inhibited crustacean muscle, focused interest on GABA as a possible inhibitory transmitter. In the first microiontophoretic tests, though GABA strongly inhibited spinal neurons, these effects were considered 'non-specific'. Strong excitation by glutamate (and other acidic amino acids) led to the same conclusion. However, their great potency and rapid actions on cortical neurons convinced other authors that these endogenous amino acids are probably synaptic transmitters. This was partly confirmed by showing that both IPSPs and GABA greatly increased Cl() conductance, their effects having similar reversal potentials. Many anticonvulsants proving to be GABA antagonists, by the 1970s GABA became widely accepted as a mediator of IPSPs. Progress was much slower for glutamate. Being generated on distant dendrites, EPSPs could not be easily compared with glutamate-induced excitation, and the search for specific antagonists was long hampered by the lack of blockers and the variety of glutamate receptors. These difficulties were gradually overcome by the application of powerful techniques, such as single channel recording, cloning receptors, as well as new pharmacological tools.