ABSTRACT When considering extracellular space and plasma membrane structure in the brain, duration of postfixation in osmium tetroxide and embedding procedure are equally as important as the dehydrating agent employed. The use of ethanol instead of acetone for dehydration does not adequately explain the finding of a 5–30 nm space instead of apposed membranes. In this investigation it was found that postfixation for 3 h stabilized the tissue sufficiently so that ethanol or acetone dehydration followed by Vestopal W, Epon 812, or Araldite 502 embedding produced similar results, and apposed, triple-layered plasma membranes were prevalent. In contrast, when postfixed for a shorter time, namely 90, 45 or 30 min, ethanol dehydration followed by Epon embedding resulted in the common ‘20-ntn ‘extracellular space and only partially triple-layered plasma membranes. The same tissue dehydrated with acetone and Vestopal embedded looked much like the equivalent 3-h postfixed material, i.e. it had apposed, triple-layered plasma membranes. When acetone dehydration and Epon embedding, or ethanol dehydration and Vestopal embedding were used with tissue of short postfixation time, the results were comparable. Apposed membranes interspersed with erratic amounts of extracellular space were common and plasma membranes tended to show an indistinct outer leaflet. Taken together the results indicate that the presence of apposed triple-layered plasma membranes is at least as valid as the common finding of a 20-nm space and partially tri-laminar membranes.