In the 1970's great strides were made in understanding the mechanism of action of amphotericin B and nystatin: the formation of transmembrane pores was clearly demonstrated in planar lipid monolayers, in multilamellar phospholipid vesicles and in Acholeplasma laidlawii cells and the importance of the presence and of the nature of the membrane sterol was analyzed. For polyene antibiotics with shorter chains, a mechanism of membrane disruption was proposed. However, recently obtained data on unilamellar vesicles have complicated the situation. It has been shown that: membranes in the gel state (which is not common in cells), even if they do not contain sterols may be made permeable by polyene antibiotics, several mechanisms may operate, simultaneously or sequentially, depending on the antibiotic/lipid ratio, the time elapsed after mixing and the mode of addition of the antibiotic, there is a rapid exchange of the antibiotic molecules between the vesicles. Although pore formation is apparently involved in the toxicity of amphotericin B and nystatin, it is not the sole factor which contributes to cell death, since K+ leakage induced by these antibiotics is separate from their lethal action. The peroxidation of membrane lipids, which has been demonstrated for erythrocytes and Candida albicans cells in the presence of amphotericin B, may play a determining role in toxicity concurrently with colloid osmotic effect. On the other hand, it has been shown that the action of polyene antibiotics on cells is not always detrimental: at sub-lethal concentrations these drugs stimulate either the activity of some membrane enzymes or cellular metabolism. In particular, some cells of the immune system are stimulated. Furthermore, polyene antibiotics may act synergistically with other drugs, such as antitumor or antifungal compounds. This may occur either by an increased incorporation of the drug, under the influence of a polyene antibiotic-induced change of membrane potential, for example, or by a direct interaction of both drugs. That fungal membranes contain ergosterol while mammalian cell membranes contain cholesterol, has generally been considered the basis for the selective toxicity of amphotericin B and nystatin for fungi. Actually, in vitro studies have not always borne out this assumption, thereby casting doubt on the use of polyene antibiotics as antifungal agents in mammalian cell culture media.(ABSTRACT TRUNCATED AT 400 WORDS)