View - Kowalewski, M. - Virginia Tech

DIETL AND KELLEY—PREDATOR-PREY ARMS RACESFIGURE 4—Constitutive, inducible (cost-benefit),and moving-target strategies. Circles representdifferent phenotypes; circle size represents growthrate in absence of predation and shading representsdegree of defense. Arrows indicate switching inpresence of predation and dashed arrows indicateswitching in absence of predation. See text forcomparison of cost-benefit (inducible) and movingtargetmodels. From Adler and Karbon (1994) withpermission of the University of Chicago Press.pressure of the inducing agent (the predator) mustbe variable and unpredictable, but sometimes strong.If the antagonist is constantly present, constitutivedefenses should evolve. 2) Availability of nonfatalpredictive cues is necessary to indicate the proximityof the threat of future attack and activate the defense.3) The defense must be effective. And 4) mostdefenses should incur direct allocation costs or othertradeoffs. The prey trade the risk of predation againstthe cost of defense.In the moving target model, changes in defense(against herbivory or predation) are induced as aform of nondirectional phenotypic escape fromadverse conditions rather than a ratcheting up ofdefenses (Fig. 4). Under this approach, preyphenotypes cannot be arrayed along a single axisfrom undefended to well-defended because theprey’s defenses are affected by a variety of extrinsicenvironmental factors other than the presence oftheir predator, each of which has an independentcomponent of effect on the predator (Adler andKarban, 1994). Consequently, there is a tradeoffin the effectiveness of different defensivephenotypes. It is predicted that prey will not fixtheir defense to one predator if that state leavesthem vulnerable to another predator. The tradeoffin the moving target case does not therefore dependon the cost of defense, but on the array of predatortypes that respond differently to different preyphenotypes (Adler and Karban, 1994). A movingtarget defense is favored if environments areunpredictable and uninformative (Adler andKarban, 1994). Thus environments that changefrequently in ecological time might not allow for aspecific inducible defense to be targeted byselection over evolutionary time.Predator responses to inducibly defended preycan br morphological, physiological, or behavioral,such as foraging movement strategies and lifehistory changes (Levin et al., 1990). Reciprocalphenotypic change has been observed in predatorpreysystems. Smith and Palmer (1994) showedthat morphology and claw strength of the predatorycrab Cancer productus was plastic. When crabswere fed mussel prey (Mytilus) with shells, theydeveloped larger and stronger claws than when fedmussels that had had their shells removed. Smithand Palmer (1994) suggested that these short-termadaptive responses to a changing environment(prey phenotype), if heritable, could produceevolutionary changes in claw size. Conversely,mussels respond to the presence of predators (e.g.,crabs) by inducing thicker shells (Leonard et al.,1999; Smith and Jennings, 2000). Agrawal (2001)argued that, because species interactions areintrinsically variable in space and time, and ifreciprocal phenotypic change is the result ofadaptive plasticity for both predator and prey, then367