The concept of gene therapy was envisioned soon after the emergence of restriction endonucleases and subcloning of mammalian genes in phage and plasmids. Over the ensuing decades, vectors were developed, including nonviral methods, integrating virus vectors (gammaretrovirus and lentivirus), and non-integrating virus vectors (adenovirus, adeno-associated virus, and herpes simplex virus vectors). Preclinical data demonstrated potential efficacy in a broad range of animal models of human diseases, but clinical efficacy in humans remained elusive in most cases, even after decades of experience in over 1000 trials. Adverse effects from gene therapy have been observed in some cases, often because of viral vectors retaining some of the pathogenic potential of the viruses upon which they are based. Later generation vectors have been developed in which the safety and/or the efficiency of gene transfer has been improved. Most recently this work has involved alterations of vector envelope or capsid proteins either by insertion of ligands to target specific receptors or by directed evolution. The disease targets for gene therapy are multiple, but the most promising data have come from monogenic disorders. As the number of potential targets for gene therapy continues to increase, and a substantial number of trials continue with both the standard and the later generation vector systems, it is hoped that a therapeutic niche for gene therapy will emerge in the coming decades.