The probability of the embolic nature of many secondary tumors was early suggested in the history of embolism. Rapidly growing tumors were known to be capable of perforating the walls of adjacent blood-vessels, especially veins, and to continue growing along the course of such vessels. The possibility of the detachment of portions of these tumors and their transfer along the course of the circulation was an inevitable inference from the results of experimentation with foreign bodies. Cancerous emboli were thus recognized as a possible variety, and their distribution was subject to the same laws as those governing emboli otherwise constituted. Multiple nodules were frequently found in the lungs in connection with tumors growing into the inferior vena cava, while multiple nodules in the liver were usually associated with tumors of the gastro-intestinal canal or other regions whose vessels formed a part of the portal circulation. The readiness with which portions may be detached after death from the soft masses projecting into the interior of veins suggests the ease with which particles may be {113} separated during life. The experiments already referred to show that isolated fragments of tissue serving as emboli may grow in the place of their reception, and it is presumable that the resulting growth takes place under the same conditions as those prevailing at the place from which the embolus started. The question whether the secondary tumor arises from the reproduction of elements transferred from the primitive disease, or whether these excite a characteristic, specific growth of the cells in the place of their retention, may still be regarded as open. The experiments favor the former view, and they alone are capable of satisfactorily determining the point in question. The secondary nodules, whatever may be their method of origin, present the peculiarities of the primitive growth. If the cells of the latter are pigmented, those of the former show the same peculiarity. If the structure of the primitive tumor contains bone, cartilage, or squamous epithelium, the secondary growths show like characters, though they may be present in the heart or other organs where such tissues are not present as normal constituents. So constant and characteristic is this feature that the structure of the tumor is usually as well displayed in the examination of the secondary as of the primitive nodule. Indeed, the structural peculiarities of the growth may be more characteristically shown in the former in those instances where the primitive tumor has undergone degenerative changes obscuring its histological features.