1. This paper reviews the evolution of the family of genes present in mammals and other vertebrates that encode gamma-aminobutyric acid (GABA) type A (GABA(A)) receptors, which are the major inhibitory neurotransmitter receptors in the central nervous system (CNS). In mammals, 16 different polypeptides (alpha1-alpha6, beta1-beta3, gamma1-gamma3, delta, epsilon, pi, and theta) have been identified, using recombinant DNA techniques, each of which is encoded by a distinct gene. The products of these genes assemble in diverse combinations to form a variety of receptor subtypes that have different sensitivities to a number of clinically relevant compounds, such as the benzodiazepines (BZs). 2. Based on a number of chromosomal mapping techniques, the majority of the GABA(A) receptor genes have been localized, in man, in four clusters on chromosomes 4, 5, 15, and the X. Furthermore, the genes that are present within these clusters have a conserved transcriptional orientation. It has, therefore, been proposed that the clusters arose largely as a consequence of two whole-genome doublings that occurred during chordate evolution, and that the ancestral cluster contained an "alpha-like," a "beta-like," and a "gamma-like" subunit gene. 3. Our laboratory has identified two additional GABA(A) receptor polypeptides (the beta4 and gamma4 subunits) in a number of vertebrate species; these do not appear to be present in mammals. We discuss here the relationship of the corresponding genes to other GABA(A) receptor genes, and conclude that their products are orthologous to the mammalian theta and epsilon subunits, respectively. 4. The GABA(A) receptor has a number of binding sites for compounds such as BZs, barbiturates, neurosteroids, and certain volatile anaesthetics. However, the only site at which endogenous compounds are thought to be active is the steroid site; this binds steroids such as certain metabolites of progesterone and deoxycorticosterone, which are synthesized in the periphery and CNS. Since the in vivo functional relevance, if any, of binding sites for other classes of compounds (such as the BZs) is unknown, the significance of differences in primary sequence, between different receptor subunits, is uncertain. We suggest that a possibly more important consequence of gene duplication is that it permitted greater flexibility in the level, pattern and regulation of expression of GABA(A) receptor genes.