View - Kowalewski, M. - Virginia Tech

DIETL AND KELLEY—PREDATOR-PREY ARMS RACESselection is a powerful force in shaping thephenotypic traits of organisms, constraints (the“spandrels” of Gould and Lewontin, 1979), whichare common and sometimes inescapable featuresof living systems, may also be important in limitingevolutionary response in an arms race.As a result of these costs, tradeoffs, andconstraints, an adaptive “stalemate” may be a morelikely outcome of an arms race than the continuousevolution predicted by the Red Queen (Rosenzweiget al., 1987; Vermeij, 1994). These limits to infiniteadaptation suggest that most of the history ofselection in an interaction may be more appropriatelymodeled with a stabilizing function, rather than thedirectional (i.e., the stronger the claw the better)function usually applied to arms races (Brodie andBrodie, 1999). However, these limits to adaptationcan be broken, leading to directional change thatcan be tracked in the fossil record—if changes inthe rules governing adaptive compromise areintroduced (see Vermeij, 1973; 1987; 1994). In thisway, selection is viewed as an episodic rather thancontinuous process (Vermeij, 1994) acting oninteractions among species.The nature of evolutionary trends predicted bythe Red Queen.—The Red Queen’s prediction oflinear (or “lock-step”; Bakker, 1983) evolutionarytrends in predator and prey has led some authors toconclude that interactions among species are notdriving evolutionary change if the expected patternis not evident in the fossil record (Stanley et al.,1983; Bakker, 1983). For example, Boucot (1990,p. 562) stated that, “after the geologically rapid,initial relation has been established [betweeninteracting species] the fossil record suggests thatthere is no subsequent, coevolutionary change, i.e.,stabilizing selection sets in.” To address this type ofskepticism, DeAngelis et al. (1984) developed acoevolutionary model of the energetics of thepredator-prey interaction between drilling naticidgastropods and their bivalve prey. In contrast toearlier models of coevolution, their model wasdeveloped with its empirical utility in mind. (Thefossil record of naticid predation is extensive andhas been key to tests of the importance of bioticinteractions in evolution—see Vermeij, 1987; Kelleyand Hansen, 1993, 1996.) Their model incorporatedan explicit potential for coevolutionary feedbackthrough size effects (Kitchell, 1986), based on sizedependentvariation in the outcome of successfulpredation (Kitchell et al., 1981). The predator wasassumed to maximize its energy intake per unit timeof foraging and the prey its allocation of energy toreproduction and defense. The models assume thatthere is a tight link between predator and prey. (Theunderlying assumption of optimality theory is thatnatural selection favors those individuals that aremost efficient in their behavior.)In the first version of the model, only theinfluence of increasing naticid predation on theallocation of bivalve energy among reproduction,overall growth in size, and shell thickness wasanalyzed (DeAngelis et al., 1985). Simulation resultsshowed that, as predation intensity increased, aninitial single bivalve defense (represented as a peakin an adaptive fitness landscape) changed to threedifferent strategies (or peaks) that varied in theamount of energy diverted to shell growth andthickness. The three alternative means of dealingwith predation are: 1) postponed reproduction(effectively running the predation gauntlet as theprey tries to grow quickly into a size refuge frompredation); 2) early reproduction coupled with someallocation of energy to thickness increase; and 3)significant allocation of energy into thickness as adefense to minimize selection by the predator(DeAngelis et al., 1985).In a later version of the model (summarized inKitchell, 1990) predator size was allowed to evolvesimultaneously with the prey traits (thusmaximizing both predator and prey fitnesses asinterdependent dynamic responses). A two-wayfeedback was thereby introduced. As traits in theprey varied to increase prey fitness, the prey in turnaffected the adaptive landscape of the predator andcaused it to change its own traits in order tomaximize its own fitness. These changes in thepredator’s traits (size), in turn, affected the prey’sadaptive landscape, causing it to adjust its ownevolutionary trajectory.Simulation results suggested that the potentialexists for both stasis and change within the dynamics357