New antibiotics are needed, (a) to control diseases that are refractory to existing ones either because of intrinsic or acquired drug resistance of the pathogen or because inhibition of the disease is difficult, at present, without damaging the host (fungal and viral diseases, and tumours), (b) for the control of plant pathogens and of invertebrates such as helminths, insects, etc., and (c) for growth promotion in intensive farming. Numerous new antibiotics are still being obtained from wild microbes, especially actinomycetes. Chemical modification of existing compounds has also had notable success. Here we explore the uses, actual and potential, of genetics to generate new antibiotics and to satisfy the ever-present need to increase yield. Yield improvement has depended in the past on mutation and selection, combined with optimization of fermentation conditions. Progress would be greatly accelerated by screening random recombinants between divergent high-yielding strains. Strain improvement may also be possible by the introduction of extra copies of genes of which the products are rate-limiting, or of genes conferring beneficial growth characteristics. Although new antibiotics can be generated by mutation, either through disturbing known biosyntheses or by activating 'silent' genes, we see more promise in interspecific recombination between strains producing different secondary metabolities, generating producers of 'hybrid' antibiotics. As with proposals for yield improvement, there are two major strategies for obtaining interesting recombinants of this kind: random recombination between appropriate strains, or the deliberate movement of particular biosynthetic abilities between strains. The development of protoplast technology in actinomycetes, fungi and bacilli has been instrumental in bringing these idealized strategies to the horizon. Protoplasts of the same or different species can be induced to fuse by polyethylene glycol. At least in intraspecific fusion of streptomyces, random and high frequency recombination follows. Protoplasts can also be used as recipients for isolated DNA, again in the presence of polyethylene glycol, so that the deliberate introduction of particular genes into production strains can be realistically envisaged. Various kinds of DNA cloning vectors are being developed to this end. Gene cloning techniques also offer rich possibilities for the analysis of the genetic control of antibiotic biosynthesis, knowledge of which is, at present, minimal. The information that should soon accrue can be expected to have profound effects on the application of genetics to industrial microbiology.