The investigation was directed to a study of the integrity of the blood-vessels in order to account for the hemorrhages. To this end the "capillary resistance test" was devised.[49] In the majority of cases this was found to be "positive" (the blood-vessels showing an increased permeability) and to become negative when antiscorbutics were given and the symptoms disappeared. This shows that the cellular structure of the vessels is altered in the course of scurvy, and indicates probably that this is an important cause of the hemorrhages. The edema of the face and ankles, the outflow of serum into the body cavities and into the muscles (Barlow) must be regarded as other evidences of the inadequacy of the vessel walls. The tendency of children with exudative diathesis to develop scurvy is perhaps still another manifestation of vascular weakness. This point of view has been strengthened recently by the pathological studies of Aschoff and Koch, who regard scurvy as a nutritional disorder in which there is a lack of some colloidal substance needed for the normal structure of the vessels. [49] A blood-pressure band, or tourniquet, is placed about the arm, and the pressure increased until the forearm becomes cyanosed and the radial pulse is almost obliterated. The pressure is then maintained at this level for 3 minutes. The principle of this test consists in subjecting the capillaries and venules to increased intra-vascular pressure to observe whether this strain results in the escape of blood. In infants the pressure was usually raised to 90 mm.; in some cases it had to be raised higher in order to entirely obstruct the return flow of the blood. The test is considered to be "positive" when the forearm shows many petechial spots. In normal infants petechiæ were almost always absent, or there were few to be seen. This is not a specific test for scurvy, but demonstrates a weakness of the vessel walls, whatsoever may be the cause. It is found to be positive in the majority of cases of scurvy. When one makes a subcutaneous puncture in infants suffering from scurvy, a small hemorrhage very often develops at the site of the puncture wound. This is not the case when one makes a hypodermic puncture in a normal person or in a hemophiliac, although it does occur in cases of purpura. This "stick test" is not a constant sign of scurvy, but, like the capillary resistance test, was found in many cases and disappeared with the subsidence of the disorder. It shows that the cells of the skin and subcutaneous tissues are affected, and possibly that their thromboplastic power is diminished.