In addition to the element of venous stagnation in producing increased transudation, the condition of the filter is of importance. The occurrence of oedema in chronic diseases, especially of the kidneys, and in those attended with protracted suppuration, continued hemorrhage, and the {71} rapid growth of tumors, has usually been attributed to the watery condition of the blood, with a diminution of the albumen. Cohnheim, however, suggests that the condition of the vessel wall is of more importance than the contents as the immediate cause of the increased transudation. The more or less protracted action of various agents--temperature, insufficient oxygen, and diminished albumen--is likely to so modify the condition of the endothelium as to favor an increased permeability of the wall. Experiments show that a simple acute hydræmia produces no increased transudation, and that a chronic hydræmia, if connected with dropsy, is likely to be influential by increasing the permeability of the wall. Even in those cases where a hydræmia and an oedema co-exist, the localization of the latter is favored by obvious disturbances of the function of the capillary walls, as in case of the cutaneous oedema after scarlatina. In like manner, a feeble heart, favoring venous stagnation, and gravitation are of importance, as general causes, in promoting dropsy in hydræmic conditions. The possibility of the occurrence of oedema through nervous influence is not to be denied. The localized and fleeting oedema of urticaria and erythema, the swollen lip and tongue in connection with digestive disturbances, are not to be explained by the two main factors of oedema--viz. venous stagnation and increased permeability of the vascular walls. Cohnheim refers to the rapid occurrence of oedema of the tongue as a result of irritation of the lingual nerve, and oedema is known to occur rapidly in cases of acute myelitis. A similar result follows the experimental destruction of the spinal cord, although the mechanism of its production is not apparent. Dropsies are subdivided, as regards their distribution, into general and local forms. The causes producing the two varieties are essentially those already described. The causes of all local dropsies are not always to be regarded as the same. Regions which are the seat of mechanical dropsies are often affected by inflammation, with abundant serous exudation--the so-called inflammatory dropsy. The properties of the effusion and exudation are quite different, the former having a small percentage of albumen, but few leucocytes, with a corresponding absence of fibrin, and few or many red blood-corpuscles. The exudation, on the contrary, is highly albuminous, though less so than the blood-plasma; it contains numerous leucocytes and much fibrin; under ordinary circumstances there are but few red blood-corpuscles.