extraordinarily successful in mice, but not in other species. The actual procedure is complicated, but the concept is simple and elegant. Rather than genetically modifying individual embryos, one at a time, simple large-scale methods are used to modify ES cells, where low efficiency is not a problem. The few cells that carry the proper genetic modification—out of the millions in culture—are then selected and turned into embryonic cells to generate genetically modified mice. The real benefit of this technique is that very low-efficiency gene insertions can be selected before the work with embryos begins. This has resulted in the use of ES cells for targeting gene constructs to selected sites in the genome3. The ability to insert a piece of DNA into a specific site in the host genome is essential for “knocking out” and possibly replacing the function of specific host genes—a technology with vast commercial importance. Although ES-like cells have been produced in other species, these have not yet been useful in generating transgenic lines of animals. Similarly to the ES cell approach, nuclear transfer has been used to turn genetically modified tissue culture cells into transgenic animals. Sheep and cow primary fibroblast cells can be genetically modified in culture; transgenic cells are then selected and, following fusion with an oocyte cytoplast, develop into genetically modified embryos and offspring4,5. This technique is much more efficient than pronuclear injection and, like the ES cell approach, may be useful for inserting genes into specific sites in the genome. Unfortunately, somatic cell nuclear transfer has not been successful with species such as the pig, and current problems with late-term pregnancy losses need to be solved. Chan et al.6 have recently reported a highly efficient system for inserting genes into unfertilized cow oocytes. This approach uses a specially designed DNA construct and a retroviral packaging and delivery system for transporting the foreign DNA into the oocyte, rather than the tedious and time-consuming microinjection approach. Retroviruses are well suited for introducing DNA into a cell, as this is how they normally infect cells. This approach appears simple and straightforward, but does not address some of the well-known limitations of retroviral mediated DNA transfer systems, such as DNA size constraints for packaging in the retroviral capsid or the introduction of retroviral sequences into the genome. Sperm-mediated DNA transfer such as that used by Perry et al. is not a new concept1. It was the subject of considerable controversy several years ago following a publication by Lavitrano et al.7 In this report, intact sperm were coated with DNA and used to fertilize eggs. Transgenic mice resulted, and the technique was hailed as a revolution in gene transfer technologies. Even with the best of efforts by several labratories, this initial success was not repeated. Perry et al. take the technique one step further. They first disrupt the sperm membrane, either by freeze-thaw or by detergent treatment. This step may facilitate the adherence of the DNA construct to the sperm surface or even allow access to the highly condensed sperm nuclear DNA. Treated sperm are then injected into an egg and give rise to transgenic mice. One might hypothesize that adhering the DNA construct to the sperm facilitates DNA transfer, overcomes problems with DNA degradation in the cytoplasm, and puts the DNA in the place where it needs to be—that is, near the decondensing sperm head at the time of male pronucleus formation. As with the other methods for gene insertion, sperm-mediated DNA transfer also has its limitations. A significant limitation is that the sperm still needs to be injected into the oocyte. Intracytoplasmic sperm injection is neither a straightforward nor a widely used technique in any species other than the human. Further work will also be required to determine if sperm-mediated DNA transfer is significantly more efficient that the standard pronuclear microinjection procedure and if it could possibly be used for targeting genes to specific sites in the genome. Significant progress has been made over the past several years in developing novel approaches for genetically modifying animals, and the work of Perry et al. is another installment in this effort. However, as was the case nearly 20 years ago, further progress is needed before we can expect to see elephant-sized cows or pig heart donors.