(5) _Traumatic Œdema._--It is not uncommon for a diffuse œdematous infiltration of the brain substance or of the arachno-pial membrane to take place in the vicinity of the injured portion of brain. This serous exude, on account of the natural adhesions of the arachno-pia, usually remains limited to the damaged area, but it may become generalised. _Mechanism._--The explanation of these widespread hæmorrhages is to be found, according to Duret, in the disturbance of the cerebro-spinal fluid which accompanies a severe blow on the head. This fluid not only surrounds the brain, but it also fills the ventricles, and permeates its substance in every direction in the peri-vascular and perilymphatic spaces. As the brain tissue is incompressible, if an area of the skull is momentarily depressed by a localised blow, space is provided for it by displacement of a quantity of cerebro-spinal fluid, which sets up a fluid wave, and this by hydrostatic pressure increases the tension of the fluid throughout the entire brain. Vessels may be lacerated at any point, either by the flow of this wave or during the ebb which follows the recoil. Hence it is that the lesion is not always at the seat of impact, but may be at the opposite side of the skull or at other remote points. [Illustration: FIG. 184.--Contusion and Laceration of Brain. Note limited lesion at point of impact on left side, and more extensive damage at point of _contre-coup_ on right. (After Sir Jonathan Hutchinson.)] _Repair._--As the disintegrated brain matter is replaced by cicatricial tissue, neither the nerve cells nor the fibres being regenerated, the loss of function of the parts destroyed is usually permanent. A localised extravasation of blood may become encapsulated, and constitute a "hæmorrhagic cyst." We have experimentally confirmed Duret's observations and agree with his conclusions. CLINICAL MANIFESTATIONS OF INJURIES TO THE BRAIN For convenience, the clinical manifestations of cerebral injury are usually described under the terms "concussion," "cerebral irritation," and "compression," but no precise pathological significance attaches to these terms, they are essentially clinical. As the conditions so described do not occur as independent entities and may overlap or merge into one another their differentiation is more or less arbitrary, and cases are frequently met with that do not run the course characteristic of any of these groups.