by a predator to lure prey (16) are rightly regarded as allomonal but are not obviously defensive. By the same token, chemicals that reduce competition for limited resources, clearly beneficial to the producer, may be defensive of those resources but are not necessarily defensive in the life of the organism producing them. Allelopathic compounds produced by a plant species may increase fitness of that plant by preempting a resource, such as water or soil nitrogen, that might otherwise be exploited by other plants (17), but in the sense that such compounds can kill potential competitors (such as nonconspecific seedlings) they are used in an offensive fashion, as for range expansion at the expense of another organism.
A defensive chemical, then, is a substance produced in order to reduce the risk of bodily harm. As such, most are poisons—defined as "any agent which, introduced (especially in small amount) into an organism, may chemically produce an injurious or deadly effect" (18). This rather restrictive definition may not be universally embraced by chemical ecologists. On one hand, the definition implies an interaction with another organism and, particularly with respect to plants, secondary compounds may fulfill many functions in the life of the producer organism other than producing injurious or deadly effects on other organisms (19, 20). Many plant secondary compounds, for example, are inducible by UV light and presumably serve to protect (or "defend") plants from damaging effects of UV exposure (21); by no stretch of the imagination can such compounds be considered poisons, since they exert no injurious effects on the damaging agent, the sun. In this context, they can no more be considered "defenses'' than cell wall constituents can be considered "defenses" against gravity. On the other hand, some investigators, while acknowledging the fact that secondary chemicals have deleterious effects on other organisms, are reluctant to ascribe their presence, particularly in plants, to selection pressure exerted by those organisms (22-24). Calling certain secondary chemicals "defenses" would be giving credence to the assertion that they exist only by virtue of the selection pressures exerted by consumers. Nonetheless, an examination of the distribution, pattern of allocation, chemical structure, and modes of action of secondary compounds in a broad cross section of organisms reveals so many striking convergences and similarities that the notion that variation in the distribution and abundance of chemicals that act as poisons results at least in part from selection by consumer organisms certainly seems tenable, if not inescapable.
One line of evidence, admittedly circumstantial, that consumers have influenced the evolution of chemical defenses is their taxonomic distri-