It is known that there are 100 Å-wide circular structures associated with the erythrocyte membrane in immune lysis. To determine whether these structures were functional holes extending through the membrane, freeze-etch electron microscopy was carried out. Sheep erythrocytes incubated with either rabbit complement or rabbit antibody (anti-sheep erythrocyte antibody) did not hemolyze and did not reveal any abnormalities in freeze-etch or negative-stain electron microscopy. Erythrocytes incubated with both complement and antibody revealed rings on the extracellular surface (etch face) of the cell membrane. Allowing for the 30 Å-thick Pt/C replica, the dimensions of the surface rings were similar to those seen by negative staining. The ring's central depression was level with the plane of the membrane; some rings were closed circles, others were crescent shaped. The cleavage face of the extracellular leaflet revealed globule aggregates, each aggregate appearing to be composed of about four fused globules. The cleavage face of the cytoplasmic leaflet was normal. When immune lysis was carried out in the presence of ferritin, ferritin was subsequently detected in all lysed erythrocytes. If ferritin was added after immune lysis was complete, only 15% of the cells were permeated by ferritin, indicating that transient openings exist in the cell membrane during immune lysis. No abnormal structures were detected when C6-deficient rabbit serum was used as a source of complement. It is concluded that antibody and complement produce surface rings, prelytic leakage of K+, colloid osmotic swelling, membrane disruption, and membrane resealing; the surface rings persist after these events.