In discussions of the ethics of human gene therapy, it has become standard to draw a distinction between the use of human gene transfer techniques to treat health problems and their use to enhance or improve normal human traits. Some dispute the normative force of this distinction by arguing that it is undercut by the legitimate medical use of human gene transfer techniques to prevent disease – such as genetic engineering to bolster immune function, improve the efficiency of DNA repair, or add cellular receptors to capture and process cholesterol. If disease prevention is a proper goal of medicine, these critics argue, and the use of gene transfer techniques to enhance human health maintenance capacities will help achieve that goal, then the “treatment/enhancement” distinction cannot define the limits of legitimate gene therapy. In this paper, I argue that a line can be drawn between prevention and enhancement for gene therapy (and thus between properly medical and nonmedical uses of gene therapy), but only if one is willing to accept two rather old-fashioned claims: 1) Some health problems are best understood as if they were entities in their own right, reifiable as processes or parts in a biological system, with at least as much ontological objectivity and theoretical significance as the functions that they inhibit. 2) Legitimate preventive genetic health care should be limited to efforts to defend people from attack by these more robust pathological entities, rather than changing their bodies to evade social injustices.