Depending on the size of the pores one wishes to produce in plasma membranes, the choice will probably fall on one of the three toxins discussed above. S. aureus alpha-toxin should be tried first when pores of 1-1.5 nm diameter are required. This is generally the case when Ca2+ and nucleotide dependence of a given process is being studied. If alpha-toxin does not work, this is probably due to the fact that the toxin either does not produce pores, or that the pores are too small. In this case, high concentrations of alpha-toxin should be tried. If this still does not work, we recommend the use of HlyA. When very large pores are to be created, e.g. for introduction of antibodies into the cells, SLO or another member of this toxin family are the agents of choice. SLO preparations need to be checked for presence of protease contaminants. Tetanolysin currently offers advantages since it is protease-free, and the size of the pores can probably be controlled by varying the toxin dose. Methods for assessing the size of pores created by such agents have been published in the recent literature, and the appropriate papers can be consulted whenever the need arises.