In sections of paraformaldehyde, fixed brain tissue, stained using immunohistochemical methods, the distribution of staining within the sections is not uniform. Whilst stained cells are seen at the top and bottom surfaces, the central thicknesses of the sections contain little or no immunoreactivity. This presents a major problem for quantification, as each section contains a population of cells that is not visualized by the staining method. Following extensive investigation of this phenomenon, we report that the failure of full thickness, immunohistochemical staining is not a failure of the immunohistochemical methodology per se, nor is it related directly to the thickness of the sections used. Rather, the problem lies in the chemistry of the tissue itself, and originates during fixation of the tissues using paraformaldehyde-based perfusion methods, which render the cell membranes impermeable to one or more components of the staining protocol. We show that this impermeability is affected by addition of membrane-disrupting agents to the fixative, and by a reduction of exposure to paraformaldehyde during fixation. The present investigation contributes to the development of new fixation protocols, optimised for use in both immunohistochemical methods and morphometric analyses. (c) 2006 Elsevier B.V. All rights reserved.