Myelin formation and its relation to the neuroglia has been a problem of continuing interest to neurocytologists since the early work of Jastrowitz (1), Boll (2), and Wlassak (3). It was not until 1928, however, that del Rio-Hortega (4) ascribed to the oligodendroglia a definite role in the process of myelinization. Since then other workers (5, 6) have supported del Rio-Hortega's thesis, although Alpers and Haymaker (6) proposed that not only the oligoglia but the astroglia are in some way implicated in the process of myelinization.Schmitt, Bear, and Palmer (7) by x-ray diffraction obtained a pattern suggesting that tile myelin of peripheral nerves is arranged as a laminated structure. This pattern has been corroborated by the electron microscopic observations of Fern~ndez-Mor~n (8), Sj/3strand (9), Geren and Raskind (10), and Robertson (11). With the resolution afforded by the electron microscope and the now available methods of fixation and thin sectioning, the mechanism of myelinization is in need of reinvestigation. Geren (12) has shown that the laminae of the myelinated sheath of a peripheral nerve are helically disposed continuations of the plasma membrane of the contiguous Schwann cell.The myelinated sheath in the central nervous system also is comprised of closely arranged lamellae and Luse (14)). The source of these membranes in the central nervous system is not obvious, for not one but many cells may abut upon a single axon. In an attempt to determine the origin of the lamellated membranes within the myelinated sheaths of axons in the central nervous system, the brains and spinal cords of young mice and rats were examined during the period of myelinization. Materials and MethodsMice and rats were killed at 1, 3, 5, and 17 days after birth. Small pieces of spinal cord, medulla, and occipital cortex were rapidly removed from decapitated animals and immedi-* This investigation has been aided in part by grants (C-2755 and B-425) from the United States

Study of the central nervous system by electron microscopyrequires identification of the neural and glial cells and of their processes. Superficial inspection. of electron micrographs reveals what first appears to be a bewildering array of components. Cells of varying sizes, shapes, and densities are recognizable by their nuclei, and in some instances, by the outline of their delimiting plasma membranes. Coursing among these cells are blood vessels, which may b e identified by the morphology of their endothelial elements and b y the presence of erythrocytes and leucocytes in their lumens. Between the cells and blood vessels, completely filling all available space, lies the neuropil. This component is a complexly interwoven meshwork of cellular processes derived from both neural and glial cells. It contains, in amounts which vary relative to one another in different locations, dendrites, axons (myelinated and unmyelinated), nerve endings, fibrous and protoplasmic processes of glial cells, and possibly~ an ill defined ground substance. By light microscopy, the relationships of: these components of the neuropil have been poorly understood, since only the larger elements could be resolved. With electron microscopy, even the smallest of the cellular processes can be visualized, but the complexity of the structures revealed makes identification of the various elements difficult.The problem of cellular identification has been approached by attempting, first, to correlate the characteristics of the neural and gilal cells as observed by light microscopy with the electron microscopic appearance of the same cells, Much information concerning the structure and interrelationships of the neural and glial cells has been obtained from use of the silver and gold chloride procedures of Ram6n y Cajal and of Del Rio-Hortega. In our interpretations, we have relied heavily on the accounts of Penfield (1) and Del Rio-Hortega (2). The nuclear differences which characterize the glial ceils can be seen both

SUMMARYThe effect of the polyene antibiotic, filipin, on the morphology of erythrocyte membranes and lipid dispersions has been investigated with the electron microscope using negative staining (phosphotungstate). Filipin induces pit formation in rat and human erythrocytes; similar pits are also produced by the antibiotic in lecithincholesterol dispersions. Pit formation requires the presence of cholesterol. Evidence is presented which suggests that the light ring surrounding each pit may consist predominately, if not exclusively, of stelol which had interacted with the antibiotic. Pits are not produced in lecithin-cholesterol dispersions by derivatives of the antibiotic which have little (perhydrofilipin) or no (irradiated filipin) hemolytic activity. Thus, the characteristic lamellar pattern of the dispersions is retained in the presence of the irradiated derivative, whereas perhydrofilipin has an effect which is apparently similar to that produced by lysolecithin. The pits produced by filipin resemble those produced in human erythrocyte membranes by immune lysis in the presence of complement but are quite different from those produced by saponin. The relevance of these observations to current concepts regarding the mode of filipin action is discussed. It is suggested that filipin may interfere with the ability of cholesterol to stabilize the bilayer configuration of phospholipids. Experiments which suggest that the sterol can function as a stabilizer are described.