Laboratory grazing experiments compared ingestion of two subclones of the dinoflagellate Alexandrium lusitanicum by gastropod veliger larvae (Nassarius sp.). While the two prey subclones originated from the same monoclonal isolate of A. lusitanicum, one possessed the ability to produce paralytic-shellfish-poisoning toxins (PSTs), while the other did not. Ingestion rates on the two Alexandrium subclones were not significantly different over a range of prey concentrations (approximately 100–660 cells ml À1), indicating that PSTs did not serve as a grazing deterrent for these larvae. However, ingestion rates on both subclones were low at the higher prey concentrations tested. Mortality of the predators also increased linearly with concentration of either subclone. These observations indicated that both A. lusitanicum subclones produced an unknown substance that inhibited and killed the grazers. Veliger mortality was not induced by culture filtrates or lysates, suggesting either that the substance was either highly labile or that contact with intact cells was required. Because toxic algae can produce multiple bioactive substances, experimental demonstrations of alleopathic effects of toxic species should not be assigned to known toxins without supporting evidence. In addition, the results show that the effectiveness of algal grazing deterrents can increase with cell concentration, which may have implications for bloom dynamics.