When Limulus sperm are induced to undergo the acrosomal reaction, a process, 50 am in length, is generated in a few seconds. This process rotates as it elongates; thus the acrosomal process literally screws through the jelly of the egg. Within the process is a bundle of filaments which before induction are coiled up inside the sperm. The filament bundle exists in three stable states in the sperm. One of the states can be isolated in pure form. It is composed of only three proteins whose molecular weights (mol wt) are 43,000, 55,000, and 95,000. The 43,000 mol wt protein is actin, based on its molecular weight, net charge, morphology, G-F transformation, and heavy meromyosin (HMM) binding. The 55,000 mol wt protein is in equimolar ratio to actin and is not tubulin, binds tenaciously to actin, and inhibits HMM binding. Evidence is presented that both the 55,000 mol wt protein and the 95,000 mol wt protein (possibly a-actinin) are also present in Limulus muscle. Presumably these proteins function in the sperm in holding the actin filaments together. Before the acrosomal reaction, the actin filaments are twisted over one another in a supercoil; when the reaction is completed, the filaments lie parallel to each other and form an actin paracrystal. This change in their packing appears to give rise to the motion of the acrosomal process and is under the control of the 55,000 mol wt protein and the 95,000 mol wt protein.Actin appears to be a ubiquitous protein in both muscle and nonmuscle cells of eucaryotic organisms (see Pollard and Weihing, 1974, for references). Myosin as well has been described in a number of nonmuscle cell types which include the amoeba, fibroblasts, Physarum, and blood platelets. Once one has mentioned the fact that actin and myosin coexist in nonmuscle cells there is almost a "Pavlovian response"--everyone wants to know where the sliding takes place, how long the myosin filaments are, etc. In fact, the sliding filament hypothesis of muscle is so firmly entrenched in our thinking that no other mechanism of motion even comes to mind, or, if it does, it is immediately dismissed as "irrelevant" or as occurring in an "oddball system." This is somewhat surprising as the myosin/actin ratio in most nonmuscle cells is very much less than that present in