The phagocytosis of glutaraldehyde-fixed horse erythrocytes by Acanthamoeba castellanii has been examined in iso-osmolal phosphate buffered saline/sucrose suspending phases of ionic strength, I, ranging from 0.17 to 0.0017. The erythrocytes were exposed, at a ratio of 15:1, to 5x10(6) amoeba in 0.2 ml volumes. The average number of erythrocytes forming a contact with an amoeba over 30 min (T(30)) was well described by T(30)=5.2 exp(-0.112xI(-0.5)). The index of the exponential 'probability of attachment' term may also be expressed in terms of either surface potential (psi(0)) or the Debye length (kappa(-1)). The probability term is formally similar to a Bolzmann factor. Electron microscopy showed that contact spreading of the amoeba over the erythrocyte took place by formation of discrete contacts and that the lateral separation distance between contacts was 0.66, 1.36 and 1.59 &mgr;m for ionic strengths 0.17, 0.052 and 0.0017, respectively. The direction of change in lateral contact separation distance was consistent with published changes in focal contact separation when amoeba move over glass or when human erythrocyte-erythrocyte adhesion occurs in different ionic strength media. The direction was also consistent with interfacial instability theory predictions for the dependence of localised membrane contact formation on interaction potential. The proportion of attached cells that were subsequently ingested correlated more strongly with the number of contacts formed along the cell-cell contact region (seam) than with the seam length at different ionic strengths.