Tuberculosis has two characteristics that are central to the therapy of disease and the prevention of spread. One, latency, or infection without clinically apparent disease, creates a large reservoir from which new cases of TB can appear and spread. The other, the slow response of active disease to treatment, creates a need for a large and expensive infrastructure to ensure completin of drug therapy. While both of these issues have long been recognized, new insights suggest possible underlying mechanisms and, perhaps, new strategies to subvert each of these problems.