and the institute of Radiology, University ofLiege (Director, Prof. Leroux) THE literature gives few details of the intimate vascularization of the optic pathway. Leber (1903) found in the optic nerve a vascular network, composed of round meshes; he described other vessels running obliquely along the whole length of the optic nerve, which became smaller with more closely woven meshes at the level of the lamina cribrosa.Wolff (1948) agreed with Leber that the distribution of the septa containing the vessels between the nervous tissue bundles reflected the vascular pattern. The branches, which derived from the very rich pial vascular network, penetrated dichotomously into the optic nerve, the septa gradually disappeared towards the intracanalicular and cranial sections, and in the centre the capillaries were less frequent and the meshwork looser (cf. also Behr, 1935).Fazio and Farina (1940) deduced from horizontal sections (not from injection) the existence of an irregular quadrangulated meshwork, orientated lengthways, built up by very small capillaries, and of equal density at the periphery and centre. Towards the optic foramen the capillaries and meshes were wider and still more irregular, and in the intracranial section a great number of very small vessels ran parallel.Eisler (1930) indicated the possibility of a network with longitudinal meshes in the sheath of the optic nerve.The subject obviously requires further investigation. TechniqueAt post mortem, the orbital contents are separated from the bone together with the periosteum, and after the superior wall of the optic foramen has been trephined, the orbital roof is removed from the inner side of the skull. The bone of the sella turcica is incised and the sphenoid bone chiselled; the internal carotid artery is taken over as long a distance as possible, and the orbit is removed complete with the optic nerve and its whole vascular supply. In the laboratory the specimen is irrigated for a few minutes; then the internal carotid artery is dissected until the opening of the ophthalmic artery is reached, or, if it is intended to inject the central retinal artery, as far as the branching of this vessel. In some specimens a branch is seen to arise from the ophthalmic artery before the central retinal artery is observed; this is described by Magitot (1908) as running to the dura mater, but in some cases (about one in four) we were able to identify it as a central artery of the optic nerve.Next a fine cannula is inserted into the artery to be injected, and without previous