Several snake venom neurotoxins are larger and more complex than the well-studied group of postsynaptic toxins exemplified by alpha-bungarotoxin. Several of these, exemplified by beta-bungarotoxin, show phospholipase A2 activity (phosphatide 2-acylhydrolase, EC 3.1.1.4) when tested in the presence of detergents. The high hemolytic activity of crotoxin, the neurotoxin of Crotalus durissus terrificus, in the presence of lecithin has been attributed to this activity. The phospholipase A2 activity of several snake venom proteins has now been compared under the physiological conditions of the hemolysis tests. It appears that only the basic component of crotoxin, B, is enzymatically active, and that its activity is not inhibited by component A under these conditions, or in the presence of deoxycholate. Phosphatidylserine is found to be digested more readily than egg white phosphatidylcholine; and also causes hemolysis in conjunction with much lower levels of crotoxin. In neither case is calcium required or stimulating. Phospholipase from Crotalus adamanteus, which is not neurotoxic, digests phosphatidylcholine more rapidly than does crotoxin, but phosphatidylserine more slowly; yet it is slightly less active than crotoxin in the hemolysis test with phosphatidylcholine, and much less with phosphatidylserine. The digestion of several phospholipids by either enzyme fails to release the expected protons in the absence of detergents at 37 degrees .beta-Bungarotoxin, highly neurotoxic, has negligible phospholipase A2 activity in the absence of detergents, and is almost nonhemolytic in conjunction with all phospholipids tested.Binding studies with (125)I-labeled compounds show that rabbit erythrocytes and ghosts have much greater affinity for crotoxin than for beta-bungarotoxin and do not bind Crotalus adamanteus phospholipase. The crotoxin complex is split in the course of binding, with only component B, the hemolytic component, becoming bound. It appears that the role of component A may be to diminish the nonspecific binding tendency of component B. Our data appear to be consistent with the concepts that affinity to membranes, particularly to specific sites on synaptic membranes, is the critical requirement for beta type neurotoxicity, and that this property, at least in some instances, has evolved from phospholipase A2 enzymes, but does not necessarily require retention and expression of enzymatic activity.