Methods for processing human and animal tissues for immunohistochemistry are reviewed. Many different fixatives and processing schemes have been suggested to be optimal for immunohistochemistry. This reflects the multiplicity of antigens that, unfortunately, makes the construction of universally adoptable procedures impossible. Fixatives can be divided into two categories: coagulant and cross-linking. Coagulant fixatives, such as ethanol, are often used to preserve immunoreactivities of large proteins such as immunoglobulins, filament proteins, and receptors. They alter the dielectric constant of proteins and precipitate them. Coagulant fixatives do not react covalently with amino acids and therefore do not change the primary structure of proteins. Also, the secondary structure may be preserved, whereas tertiary structure (protein folding) may be changed. Lack of effects on primary and secondary protein structure may explain why coagulant fixatives preserve many protein epitopes. Low-molecular-weight material will often be extracted during coagulant fixation. Several dislocation artifacts have also been detected during coagulant fixation, and tissue shrinkage may present a problem. Aldehyde fixatives, such as formaldehyde and glutaraldehyde, preserve structure and immobilize antigens to a much greater degree than do coagulant fixatives. Cross-linking fixatives attack the primary structure of proteins and form cross-links and adducts that often involve amino, sulphydryl, and amide groups. This results in a much more profound effect on primary, secondary, and tertiary protein structure, but is necessary for optimized morphology and for immobilization of soluble antigens at their native sites. Formalin is by far the most commonly used cross-linking fixative for immunohistochemistry. The fact that formalin fixation is no single well-defined procedure is stressed, and the need for standardization in fixative concentration, time, temperature, pH, and osmolarity is emphasized. Although many antigenic structures are lost during formalin fixation, our studies show that it often is the combination of formalin and paraffin embedding that is the real culprit. Mild formalin fixation may thus be optimal for some sensitive antigens, but subsequent processing may result in extractions, superimposed ethanol fixation, and heat modifications. Thus, formalin fixation followed by cryo-protection and cryostat sectioning preserves a larger array of antigens than do formalin-paraffin procedures. Although it often is stated that a particular fixation preserves or destroys a certain antigen, this is not necessarily true. Thus, antibodies recognize only very small parts of antigens, and such epitopes may be affected differently by fixation. Thus, it has been possible to find “formalin-resistant antibodies” to many antigens that also work well on archival material. In other cases, chemical or enzymatic demasking procedures have successfully retrieved antigens that were “lost.” The bottom line, however, is still that tissue preparation is based mainly on trial and error and that no golden rules can be offered.