There is ample circumstantial evidence from observation of the natural history of tuberculosis in humans and experimental animals that Mycobacterium tuberculosis is capable of adapting to prolonged periods of dormancy in tissues, and that these dormant bacilli are responsible for latency of the disease itself. Furthermore, the dormant bacilli are resistant to killing by antimycobacterial agents. A systematic evaluation of the mechanism of dormancy, and of attempts to abrogate latency will require a better understanding of the physiologic events that attend the shiftdown into dormancy. There are probably two or more stages in the shiftdown of Mycobacterium tuberculosis from active replication to dormancy as bacilli in unagitated cultures settle through a self-generated O2 gradient into a sediment where O2 is severely limited. One step involves a shift from rapid to slow replication. The other involves complete shutdown of replication, but not death. Presumably this last step includes completion of a round of DNA synthesis. The shiftup on resumption of aeration includes at least three discrete sequential steps, the production of RNA, the ensuing synchronized cell division and, finally, the initiation of a new round of synthesis of DNA. Three markers of the process of shiftdown of Mycobacterium tuberculosis to dormancy have been described, namely the changes in tolerance to anaerobiosis, the production of a unique antigen and the ten-fold increase in glycine dehydrogenase production. Additional markers represented in the shiftup and shiftdown process may yet be discovered, and determination of their specific functions should provide insights into the mechanisms of dormancy and latency in tuberculosis, and into strategies for preventing reactivation of the bacilli and development of disease.