The traumatic[2] theory. There is much in favor of this. In a certain number of cases tumors do develop at the site of injuries. The coincidence of injury and tumor is apt to be overestimated because of the strong tendency to connect succeeding events. Tumors are not most common on those parts of the body which are most exposed to injury. They are rare, for instance, on the hands and feet, and very rarely do they appear at the site of wounds caused by surgical operations. For those tumors which develop in intra-uterine life it is difficult to assign injury as a cause. There does, however, seem to be a relation between tumors and injuries of a certain character. The natives of Cashmere use in winter for purposes of heat a small charcoal stove which they bind on the front of the body; burns often result and tumors not infrequently develop at the site of such burns. Injuries of tissue which are produced by the X-ray not infrequently result in tumor formation and years may elapse between the receipt of the injury and the development of the tumor. These X-ray injuries are of a peculiar character, their nature but imperfectly understood, and the injured tissues seem to have lost the capacity for perfect repair. In regard to the possible action of both injuries and parasites in causing tumors, the possibility that their effects on different individuals may not be the same must be considered. In addition to the trauma or the parasite which may be considered as extrinsic factors, there may be conditions of the body, intrinsic factors, which favor their action in tumor development. The peculiar tissue growth within the uterus called decidua, which occurs normally in pregnancy and serves to fasten the developing ovum to the inner lining of the uterus, may be produced experimentally. This growth depends upon two factors, an internal secretion derived from the ovary and the introduction into the uterus of a foreign body of some sort; in the case of pregnancy the developing embryo acts as the foreign body. It is not impossible that some variation in the complex relations which determine normal growth may be one factor, possibly the most important, in tumor formation.