In the history of life, species have adapted to their consumers by evolving a wide variety of defenses. By contrast, animal species harvested in the wild by humans have not adapted structurally. Nonhuman predators have high failure rates at one or more stages of an attack, indicating that victim species have spatial refuges or phenotypic defenses that permit further functional improvement. A new compilation confirms that species in the wild cannot achieve immunity from human predation with structural defenses. The only remaining options are to become undesirable or to live in or escape to places where harvesting by people is curtailed. Escalation between prey defenses and predators’ weapons may be restricted under human dominance to interactions involving those low-level predators that have benefited from human overexploitation of top consumers.

Almost all ecosystems of the last 500 million years have supported consumers, animals that kill and eat other multicellular organisms. Potential victim species have adapted to their enemies by evolving behavioral, structural, or chemical defenses and by adjusting life-history traits even as the consumers themselves have on average become more powerful (Vermeij 2004). As the most powerful consumers that have ever lived, humans should have elicited morphological adaptive responses in the many species we harvest for food or ornament in the wild; yet, there is a striking absence of structural adaptation in the undomesticated animal species we hunt, fish, and gather. Where adaptation has been demonstrated or inferred, it is either behavioral, as manifested by increased fear, vigilance (Darwin 1859), or demographic, as reflected in trends toward smaller adult size, earlier maturation, and a faster life cycle (Conover and Munch 2002; Olsen et al. 2004; Birkeland and Dayton 2005; Conover et al. 2005; Hutchings and Baum 2005; Kuparinen and Merilä 2007; Fenberg and Roy 2008; Darimont et al. 2009). Under this new regime, which Allendorf and Hard (2009) call “unnatural selection,” many traits that contributed to effective adaptation against nonhuman enemies have lost their usefulness or even become liabilities. In other words, characteristics such as large adult size, a well-armored shell in molluscs, elaborate antlers or horns in large herbivorous mammals, and large meaty claws in crabs and lobsters were effective in defense, competition, and mate choice before humans appeared on the scene; but these traits attracted humans and therefore became disadvantageous under the new human-dominated selective regime.

What accounts for this apparent lack of structural adaptation to us as predators? Has there been insufficient time or heritable variation for natural selection to do its work, or is there something about the way we select prey that prevents prey species from enhancing old defenses or establishing new ones? Are there failures of antipredatory defense in nature and, if so, what can they tell us about the conditions that are necessary for the evolution of deterrents?

I suggest that the elimination of size refuges for most of the undomesticated animal species exploited by humans has closed all pathways of structural defensive adaptation, and has left very few adaptive options open. Using shore animals as examples, I propose that the radical change in the criteria for antipredatory success stemming from the evolution of the human species has effectively restricted the predator–prey arms race to interactions involving predatory species that have benefited from the human overexploitation of apex predators.

Adaptation to NonHuman Predators

In an analysis of 132 prey categories consumed by 97 predatory species ranging from rotifers and insects to fish, birds, and mammals, I showed that the postdetection probability of successful capture and consumption exceeded 90% in only 22 cases (16%) (Vermeij 1982a). These high success rates were achieved for small, slow, poorly defended prey species or stages. Even if small size classes of prey suffered a near-certain chance of being killed once captured, large individuals stood a better chance of surviving an attack (Vermeij 1982a).

From the prey's perspective, the high incidence of predator failure means that most individuals of victim species possess traits sufficient to protect them from predation. These traits serve as the basis for adaptive elaboration in the face of predator evolution as long as the criteria for success remain unchanged. Even if a few powerful predators can kill the best-defended prey, many predators of the same species are smaller and less potent and therefore fail in their attacks. As long as failure is common, the potential exists for prey species to persist and to evolve more effective defenses of the same kind that worked before. Nearly all morphological defenses that function primarily at the subjugation (or resistance) phase of an attack work better in large animals than in small ones. Prey species relying chiefly on resistance defenses therefore have at least a partial size refuge from predation. Once they reach the size range of partial or complete invulnerability, prey species are freed from the predatory constraint and can devote more resources to competition and reproduction.

A refuge in large size is achievable because predators are limited by the size, gape, and generated power of feeding structures. Even the largest predators that hunt in groups—a habit associated with higher success rates of predation (Kruuk 1972; Schaller 1972; Enders 1975; Mech 1976; Griffiths 1980; Hector 1986)—do not escape this limitation. For example, large marine mammals overwhelmingly take prey less than 30 cm in length despite the availability of larger potential victims (Etnier and Fowler 2010). Springer and colleagues (2003) compellingly argue that the high incidence of sublethal injuries on large whales represent failed attacks by killer whales (Orcinus orca), socially hunting top predators in the ocean today. Powerful legs for kicking and horns for goring predators protect large mammals such as moose (Alces alces) and African buffalo (Syncerus caffer) against such top terrestrial predators as wolves (Canis lupus), lion (Panthera leo), and hyena (Crocuta crocuta) (Kruuk 1972; Schaller 1972; Mech 1976). Although predators capable of dismembering prey outside the predator's body or hunting socially can bring down victims larger than they are, large terrestrial herbivores exceed the size of the largest predators by a factor of 10 (Burness et al. 2001), a situation that has existed ever since predatory tetrapods capable of attacking large herbivores evolved in the Early Permian (Bakker 1977). Although this size advantage apparently does not hold in fresh water or marine environments, where predators have been the largest animals since Early Cambrian time (Vermeij 2004), species with passive defenses such as armor have enjoyed structural refuges from their predatory enemies throughout Phanerozoic history (Vermeij 1987, 2004).

In the absence of effective resistance and a refuge in large body size, species rely on traits that render individuals unrecognizable or undesirable as prey, promote escape before capture, or enhance fecundity in the face of heavy losses of undefended victims. Still other species are restricted to safe environments where predators are absent or ineffective, as is the case underground, in the air, in physiologically challenging environments, and on or in the bodies of well-defended hosts (Vermeij 2004). Species with these responses tend to be small bodied.

On a geological time scale, major shifts in the criteria for success against enemies have occurred many times as consumers with new or much more powerful means of exploitation evolved and as old defensive syndromes became obsolete. For example, land plants and seaweeds with growth zones near the base of stems and leaves (or blades) rather than at apices became the norm in many ecosystems, notably grasslands and kelp forests, when herbivorous mammals with high metabolic demands and large appetites evolved (Vermeij 2004, 2010). Passive armor has been replaced to more active defenses in many animal groups when faster, more powerful predators evolved. In these cases where one adaptive syndrome was replaced by another, adaptive options remained available even as consumers with higher performance levels arose and diversified. Good examples are to be found among cephalopods, in which lineages relying on passive shell armor were replaced by lineages emphasizing speed, aggression, venom, or toxicity in the face of intense competition and predation by fish and cetaceans (Packard 1972; Vermeij 1987, 2004).

Humans as Predators

The spread of modern humans represents one of the great ecological and evolutionary transformations in the history of life. As culture amplified or replaced genes as the primary medium of change in our species, humans achieved global economic dominance in the biosphere (Vermeij and Leigh 2011). Initially, Homo sapiens and its immediate ancestors were largely terrestrial omnivores, but intertidal resources were already being used by Neanderthals as early as 200,000 years ago (Morales-Muniz and Roselló-Izquierdo 2008) and 164,000 years ago by modern humans in South Africa (Marean 2011). Intertidal animals appear to have been the chief food sources for prehistoric populations in South Africa, western South America, California, and at times in the North Sea Basin (Jerardino et al. 1992; Siegfried et al. 1994; Richards et al. 2003; Klein et al. 2004; Reitz et al. 2008; Rick et al. 2008; Erlandson et al. 2009), and have always been important in Oceania and the West Indies (Spennemann 1987; Thomas 2007; Anderson 2008; Fitzpatrick et al. 2008; Jones 2009; Ulm 2011). Exploitation of fish and marine mammals came later (Klein et al. 2004; Erlandson et al. 2009; Marean 2011), and agriculture came later still.

There is strong evidence that humans have preferentially harvested large species, and large individuals of smaller species, in the wild. Fenberg and Roy (2008) document 108 living species in which large adults are at greatest risk of human exploitation. These species range from terrestrial mammals to marine and freshwater fish and marine invertebrates, and include animals harvested for food as well as species collected for ornamentation and other purposes (Roy et al. 2003; Teske et al. 2007).

I expanded Fenberg and Roy's (2008) survey to all studies of human exploitation of marine molluscs and echinoderms, I was able to find in the archaeological and ecological literature. For each species in which exploitation relative to body size was assessed, I ascertained whether harvesting targeted large individuals (coded +) or was unselective with respect to size (coded 0). Preferential exploitation of large individuals was inferred when the original source indicated statistical significance of the results. I made no attempt to estimate the strength of selection, because data were generally unsuitable for this purpose and in any case refer to a phenotype whose precise genetic contribution was not known.

My survey of shore animals (Table 1) covers 30 studies and 40 species. It reveals 35 species (87.5% of the total) in which humans select for large individuals, five species (12.5%) in which size-selective foraging was not found, and no species in which humans targeted small individuals. This strong preference for large prey is worldwide, and has been evident in prehistoric hunter gatherers as well as among present-day foragers. For species that live in both the intertidal zone and below the low water line, the largest individuals typically live in subtidal environments, where they occupy a partial refuge in space from human predation. If subtidal populations are large enough, nonhuman predators there may exert selection strong enough to counterbalance human-caused natural selection against large body size. For strictly intertidal species such as many limpets (species of Cymbula, Fissurella, Lottia, Patella, and Siphonaria) and a few other gastropods (Haliotis cracherodi, Osilinus lineatus, and Acanthinucella spirata), there are no subtidal refuges, and only some populations on inaccessible wave-exposed shores may be immune from human predation. Insofar as body size is heritable, these species may be undergoing strong selection against large size and in favor of maturation that occurs earlier and at smaller sizes.

Table 1.  Shore invertebrates exploited in the wild by humans.
Size bias?Category and taxon Abalone:LocationReference(s)
  1. + Preferential harvesting of large individuals; 0 indicates no size bias.

+Haliotis cracherodiiCalifornia(Raab 1992; Braje et al. 2007)
+Haliotis idaeSouth Africa(Branch and Moreno 1994)
+Haliotis rufescensCalifornia(Braje et al. 2007)
+ Cymbula argenvilleiSouth Africa(Klein et al. 2004)
+ Cymbula granatinaSouth Africa(Klein et al. 2004; Sealy and Galimbert 2011)
+ Cymbula granularisSouth Africa(Klein et al. 2004)
+ Cymbula oculusSouth Africa(Branch and Odendaal 2003)
+ Fissurella limbataChile(Jerardino et al. 1992)
+ Fissurella volcanoCalifornia(Roy et al. 2003)
+ Helcion concolorSouth Africa(Branch 1975)
+ Lottia giganteaBaja California and California(Pombo and Escofet 1996; Roy et al. 2003)
+ Patella candeiCanary Islands(Hockey 1987)
+ Patella vulgataNorthwestern Spain(Gutiérrez-Zugasti 2011)
 Coiled herbivorous gastropods:  
+ Agathistoma aureotinctaCalifornia(Roy et al. 2003)
+ Cittarium picaWest Indies(Fitzpatrick et al. 2008)
0 Conomurex luhuanusKiribati(Thomas 2007)
0 Gibberulus gibberulusLau Islands (Fiji)(Jones 2009)
+ Lobatus gigasTurks and Caicos(Stager and Chen 1996)
+ Osilinus lineatusEngland(Mannino and Thomas 2002)
+ Turbo sarmaticusSouth Africa(Klein et al. 2004; Sealy and Galimbert 2011)
+ Turbo spp.Lau Islands (Fiji)(Jones 2009)
 Predatory gastropods:  
+ Concholepas concholepasChile(Jerardino et al. 1992; Castilla et al. 1994)
+ Nassarius kraussianusSouth Africa(Teske et al. 2007)
+ Crassostrea virginicaMaryland(Kent 1988)
+ Ostrea edulisDenmark(Bailey et al. 2008)
0 Choromytilus meridionalisSouth Africa(Klein et al. 2004; Marean 2011)
+ Mytilus californianusCalifornia(Braje et al. 2007)
+ Perna canaliculusNew Zealand(Anderson 2008)
 Burrowing bivalves:  
+ Anadara antiquataTongatabu(Spennemann 1987)
0 Anadara uropigimelanaKiribati(Thomas 2007)
+ Austrovenus stutchburyiNew Zealand(Anderson 2008)
+ Cerastoderma eduleDenmark(Bailey et al. 2008)
+ Gafrarium spp.Tongatabu(Spennemann 1987)
+ Paphies australisNew Zealand(Anderson 2008)
+ Senilia senilisSierra Leone(Okera 1976)
 Sea urchins:  
+ Evechinus chloroticusNew Zealand(Anderson 2008)
+ Strongylocentrotus spp.Aleutian Islands(Corbett et al. 2008)

The overwhelming tendency of humans to take the largest individuals in exploited marine molluscan populations implies that the success rate of predation by humans is very high. Although I have been unable to find direct evidence for this conjecture, the low abundance of harvested species in unprotected areas as compared to nearby reserves and the endangered status of many overfished species strongly indicate that few individuals in exploited habitats remain immune to human predation. Direct data on success rates of humans as predators compared to other predatory species would be most welcome, and could be acquired by field anthropologists by following individual foragers and evaluating which potential prey are taken and which remain uncollected.

With the elimination of a size refuge for many of the species humans harvest in the wild, the morphological adaptations associated with large size are no longer effective. In the absence of structural defenses that could serve as the basis for enhanced protection in the novel human-dominated evolutionary environment, many of the species we harvest in the wild lack the phenotypic (and presumably the genetic) variation required for further structural adaptation. The options for evolving resistance defenses are therefore largely closed to human-exploited animal species without purposeful human selection and breeding, which implies domestication and farming.

One possible exception among shore animals is the ability to cement the external skeleton to rock. Spennemann (1987) noted that cemented bivalves of the genera Chama and Spondylus are infrequently harvested in Tonga because of the difficulty of collecting them; but oysters, whose valves are also cemented to rocks or to each other, are favorite targets throughout the world (Kent 1988; Kirby 2004; Kirby and Miller 2005). Tools like picks and knives would in any case suffice to overcome the challenges of collecting cemented bivalves and barnacles.

Besides the behavioral and life-history traits that are still available to the highly mobile species we fish or hunt, the most accessible adaptive pathway under the human selective regime is to become undesirable as food or ornament. Toxicity is one way to achieve this, particularly for immobile or slow-moving species (Branch and Moreno 1994; Hockey 1994). The omnivorous Chilean sea urchin Tetrapygus niger may be avoided by human foragers for this reason, although the causes of its undesirability have not been investigated (Hockey 1994).

The evolution of toxicity in animals can proceed in three ways: (1) in-house synthesis of toxins; (2) the incorporation and sequestration of deterrents from food; and (3) symbiosis with toxin-producing microbes. In-house synthesis appears to be rare, but does occur in many echinoderms and in some opisthobranch molluscs (Cimino and Ghiselin 2009) as well as in plants, algae, fungi, and bacteria. Siphonariid limpets have glandular secretions (Pinchuck and Hodgson 2009) that offer some protection against nonhuman predators (McQuaid et al. 1999) and that give the large tropical eastern Pacific S. gigas a bitter taste (Morrison 1963). Despite these chemical defenses, S. gigas is extensively harvested for food by people in Costa Rica (Ortega 1987); and I have observed people in Ghana scraping rocks to remove the much smaller S. pectinata for food.

Diet-induced acute toxicity is widespread among potentially edible marine species and presents a significant deterrent to humans. When eaten, reef fish that consumed epiphytic dinoflagellates or that preyed upon animals that fed on those dinoflagellates cause lethal ciguatera in humans (Lobel et al. 1988; Kohler and Kohler 1992). Herbivorous and carnivorous crabs on reefs may contain dinoflagellate-derived toxins and cause death when eaten (Llewellyn and Endean 1989; Tsai et al. 1997). Planktonic dinoflagellates are responsible for paralytic shellfish poisoning when humans consume infected bivalves. Although some species seem very frequently to be toxic, such as the reef crab Zosimus aeneus, none is obligately so, and there is considerable regional and seasonal variation in the composition and concentration of toxins (Tsai et al. 1997).

Among suspension feeders, most species appear to be palatable under normal conditions. The only possible exceptions are rhynchonelliform (articulate) brachiopods, which are said to be less preferred by most predators than bivalves and have therefore been considered unpalatable (Thayer 1985; Thayer and Allmon 1991); but, Harper (2011) has pointed out that predation is a widespread cause of death of living brachiopods, and any feeding deterrents in the group remain unidentified. Unless suspension feeders can evolve symbioses with toxic microbes, consistent chemical defense against humans is an unlikely solution to human predation.

A general limitation of toxicity in animals is that the poisons tend to be concentrated or sequestered in particular parts of the body. Skin (or exoskeleton) and the liver are particularly prone to be toxic, and therefore can be removed by knowledgeable human consumers. It should be noted, however, that most parasites and pathogens that humans have tried to eradicate, sometimes successfully, depend on rapid chemical adaptation to counter human intervention and the immune system.

Moreover, my impression is that toxic animals tend to be small, relying on predators tasting and then rejecting them. Larger prey that require a greater subjugation effort by predators tend not to be toxic. Surprisingly, data on the size distribution of prey species with different classes of defense—camouflage, armor, toxicity, speed, and aggression—have not been systematically collected. Investigations into the size distribution of antipredatory adaptive syndromes would be highly rewarding for a deeper understanding of evolutionary predator–prey dynamics.

As many authors have pointed out, the declines in size and productivity observed in many of the species we harvest in the wild can be reversed only by imposing maximum size limits on the individuals taken (Conover and Munch 2002; Birkeland and Dayton 2005; Conover et al. 2005). Besides enabling prey species to be more productive and more sustainably harvestable, maximum size limits would also give the targeted species more scope to respond to their nonhuman predators and competitors, some of which will have increased in abundance because of release from predation by large apex predators that human exploitation has all but eliminated. For example, the collapse of cod (Gadus morhua), an apex predator in the northwestern Atlantic, has led to a huge increase in abundance of shrimps (Worm and Myers 2003), lobsters, and crabs (Steneck et al. 2004; Butler et al. 2006; Bourque et al. 2008). Observed 20th-century increases in predation-induced shell repair in the northwestern Atlantic shore gastropods Littorea littorea (Vermeij 1982a) and Nucella lapillus (Vermeij 1982c) and phenotypic increases in shell thickness in L. obtusata and N. lapillus (Vermeij 1982b, c; Seeley 1986; Trussell and Smith 2000; Rochette et al. 2007; Edgell et al. 2008; Fisher et al. 2009) have generally been interpreted as responses to the spread of the introduced European green crab (Carcinus maenas), but these changes may also have resulted from an overall increase in the abundance of native shell-crushing crabs (Cancer spp.) and lobsters (Homarus americanus) as the predators (including cod) of these crustaceans were overfished. Moreover, warming may also have contributed to the increase in shell crushing and body size in the western Atlantic populations of N. lapillus (Fisher et al. 2009).

These indirect effects of humans on the selective regime of species we do not harvest in large numbers are likely to be extremely widespread. By effectively eliminating large apex predators, which themselves eat large secondary predators, humans have increased the abundance of the secondary predators, many of which have broad diets emphasizing small, slow, armored herbivores and suspension feeders. Aside from the northwestern Atlantic cases discussed in the preceding paragraph, human-induced alterations in the selective regimes of species other than those we target for harvest in the wild have scarcely been studied. Whether and how species can evolve and adapt under human hegemony remains one of the outstanding and ignored questions for our future.

In prehuman times, the most potent predators and competitors exercised both direct and indirect evolutionary controls on victim species by regulating abundances and acting as selective agents. In the new world order, in which humans dominate the planet ecologically, a direct role for humans as selective agents for adaptations other than those related to life history is limited to the species we domesticate. Escalation of prey defenses and predators’ weapons in the future is likely to be restricted to interactions in which predators have been ecologically released by human elimination of species that previously functioned as top consumers. Ecologists and evolutionary biologists have barely begun to probe the nature and implications of this cataclysmic shift in the regime of resources and selection to which species on Earth live now and in the future.

Associate Editor: M. Hart


I thank C. W. Marean for asking me the question that prompted this paper, J. J. Stachowicz and B. Worm for pointing me to some helpful references, and T. Michlin for technical assistance.