1. It must be accepted from a comparison between the histological changes described in the focal necroses due to abrin and ricin, diphtheria, eclampsia, blood-serum intoxication, typhoid fever, lobar pneumonia and glanders, and those described here in three cases of tuberculosis, that the last are also instances of focal necrosis. 2. It may be claimed that the areas of necrosis in the three cases of tuberculosis had their origin in ischæmia, for the reason that they were numerous in the spleen and were not found in the kidney. Against such an explanation of their causation are the facts: that other forms of necrosis affect frequently the spleen and seldom the kidney; that thrombosis of vessels or any fixed relation of the areas in the spleen to blood-vessels could not be established; that the tubercle bacilli were invariably present; and lastly and of no slight importance, that these necroses were all present in cases of marked marasmus. 3. That the areas of necrosis in these three cases might be phenomena resulting from a mixed infection with pyogenic bacteria is possible, but no groups, colonies or even single bodies could be found which answered to the requirements in shape or staining properties for such bacteria. Even were it the case that the focal necroses here presented are the histological lesions of a septicæmia occurring as a complication of tuberculosis, the presence of the bacillus of tuberculosis in the areas of necrosis would still require explanation. 4. Lastly, it is not unreasonable to suppose that a diminished resistance on the part of the tissues incidental to marasmus is an important factor in the production of these necrotic lesions.