In December 2000, Janice and I traveled to Queensland, Australia, to study antipredator behavior in rock wallabies and birds. The birds of the Atherton Tablelands are especially diverse, and I was excited to increase the number of unique species we studied. I had just started to develop a large comparative database of flight-initiation distance in birds, a subject not well studied at the time. Flight-initiation distance, or FID, is the distance at which an animal begins to flee an approaching object. Virtually all species respond fearfully to approaching humans, and there’s now a very large literature on “flushing,” or encouraging animals to flee from their current location. Janice and I were creating a large database of human-stimulated FIDs to answer the fundamental question of why some but not all species tolerate humans. By comparing various species’ fearful responses, we aimed to understand how these relationships had evolved, creating what we describe as an ecology of fear.
Data collection was particularly enjoyable because we began by finding and identifying the local birds. Once we spotted a bird, we slowly walked toward it while counting measured paces, noted the location at which it looked around, the location at which it fled, and the location where it was when we began our approach. Our vigilance and diligence paid off: a fortnight later we left with a much larger data set and only a few scabs from the abundant terrestrial leaches. What did we learn from such apparently simple data? Escape decisions nicely illustrate the economic logic that underpins all decision making—whether by animals or by us.
How do animals assess risk? In this chapter we will explore the huge literature on FID to understand the dynamics of risk assessment. These lessons have important implications for ecotourism since animals respond fearfully to us, just as they respond fearfully to predators. By understanding the economics of fear we gain more insights into the trade-offs animals (and humans) make on a daily basis that enable them (and us) to live another day.
Animals that allocate only the minimum time, energy, and resources needed when responding to threats will outcompete and ultimately leave more descendants than those that overreact to fearful situations. The common assumption that animals should flee approaching predators as soon as they detect them may be wrong. In fact, Darwin, writing in what’s now known as the Voyage of the Beagle, was puzzled by why he could get so close to animals on the Galápagos, musing about what later became known as island tameness. Since Darwin’s observations, scientists, especially behavioral ecologists, look for rules that explain behavioral diversity—both within-species diversity and between-species diversity—in an explicitly functional and evolutionary context. Behaviors that have net benefits are selected and evolve; those that have net costs are eliminated by natural selection.
Ultimately, costs and benefits should be quantified in terms of fitness. Behavioral ecologists think of “fitness” a bit differently. We don’t measure how many daily steps an individual takes but rather how many genes from an individual survive to populate future generations. After all, evolution is about having your genes survive at higher frequencies than other individuals’ genes. Viewed this way, fit individuals are those whose genes are disproportionately represented in future generations, while unfit individuals are those not genetically represented in future generations.
Though theoretically straightforward, it’s exceedingly difficult to properly measure evolutionary fitness in most situations. It’s not so easy to follow genes into future generations in the field, nor to attribute a specific action to a fitness outcome. Thus, behavioral ecologists take a shortcut: they quantify things that should be correlated with fitness: time, energy, and opportunity. Viewed this way, a costly behavior is one that takes time or energy, or prevents an individual from doing something else that may offer better results. All else being equal, it’s more costly to run than to walk and less costly to sit than to move. Expending energy is a bad thing, unless it’s necessary for survival. Running or fleeing from predators, for example, is clearly worth the cost. Through use of these fitness surrogates and through FID experiments like the ones that Janice and I conducted in the Atherton Tablelands, we can identify rules that govern escape behavior. Let’s take a walk through some of those rules, and I’ll show you how they describe evolutionary fitness.
In 1986, Ron Ydenberg and Larry Dill, two researchers based at Simon Fraser University in suburban Vancouver, published a deceptively simple and remarkably elegant model of escape behavior. They realized that it was essential to consider not only the benefits of flight—fleeing immediately upon detecting a predator—but also its costs. If an individual flees too soon, an opportunity is lost. If an individual flees too late, the predator may be successful.
To illustrate this, they graphed the relationship between the distance to the predator and the cost of remaining, which is their term for the benefits of fleeing. At large distances, the cost of remaining is relatively low. It increases as the hypothetical predator gets closer and closer to the prey. Thus, if distance is the x-axis on a graph, the curve slopes downward as distance increases. On the same graph, they also plotted the relationship between predator distance and the cost of fleeing, which increases linearly. The cost of fleeing from a distant predator is much higher than the cost of fleeing from one nearby. Because the cost-of-remaining curve declines with distance while the cost of flight increases, there is a point where the lines cross, and this point represents the optimal distance, from the prey’s point of view, for flight.
What sorts of things could influence the cost of fleeing? Wouldn’t it be better to flee immediately if there is a predator nearby? Ron and Larry suggested that lost foraging opportunities are one such cost. Envision two birds: one is foraging and the other is simply sitting and digesting between bouts of foraging. If foraging is at all limited, then the bird that is foraging would lose more by stopping its activity to flee an approaching human. Thus, it would have a greater cost of flight. Using economic logic, Ron and Larry noted that the optimal location to flee was farther away when the cost of flight increased more slowly (such as for a resting animal) and closer when the cost of flight increased more rapidly (such as for a foraging one). In other words, animals should flee when the benefits from fleeing exceed the costs of remaining.
More generally, behavioral ecologists consider behavior optimal when an individual selects the strategy that produces the greatest fitness when given a set of possible strategies, whether immediately fleeing when a predator is detected, or waiting ten seconds before fleeing, or fleeing when the benefits most exceed the costs. Since we’ve defined maximizing fitness as maximizing the benefit-to-cost ratio, this decision should, by definition and on average, be the optimal solution. We assume that animals who flee optimally will survive and allocate their energy into producing more descendants than individuals using the other possible strategies. Importantly, “optimal” is defined with respect to possible strategies—not the best conceivable, but rather the best in a set of possibilities. While it might be optimal for an impala to fly away from a pack of wild dogs, impala can’t fly.
Once Ron and Larry’s model introduced the idea that escape decisions should be optimized, many researchers began to study factors that influence escape behavior. Some studies found that animals in larger groups tolerated closer approaches before fleeing. Other studies found the opposite. Some studies found that when animals were in dense cover, they tolerated approaches to a closer distance. Other studies found the opposite. What accounts for this diversity of results in experimental studies? And with this diversity, is there a general principle that explains the many paths of escape behavior that we see in nature?
Many disciplines of scientific research borrow methods and ideas from other fields, and old problems are often solved with new tools and techniques. Behavioral ecology is one such discipline. We have had much success using ideas from other disciplines to help us identify general principles. Economic tools can help us understand animals’ foraging decisions, and similar tools can give us insight into when it’s profitable to defend a territory and when it’s profitable to fight for access to resources.
Species share relationships with their close relatives. Dogs and wolves, which are closely related, are more similar to each other in their behavior than are dogs and cats, which are more distantly related. Closely related species may share body size, brain size, eye size, reproductive strategies, habitat preferences, and overall appearance. If these different traits influence escape behavior, then we’ll want to know how to account for the similarity that is expected because some species in a given data set may be more closely related than others. This can be a problem whenever there are biases in how data are collected. If, for instance, it’s very easy to collect data on different species of crows, we wouldn’t want a disproportionate number of crows to influence our general conclusions about the effect of, for example, body mass or group size on escape behavior in all bird species. Evolutionary biologists and statisticians have developed a number of methods that permit us to remove the effects of these close phylogenetic relationships and truly isolate the relationship between body mass or group size on escape behavior. Yet different comparative studies might produce different answers because they included different species that were studied at different times and in different locations. After all, we expect the environment to factor into some of the differences in behavior, due to how species have adapted to their environment.
Fortunately, there is a statistical approach that permits us to combine the results of different studies and allows us to draw more general conclusions about the diversity of behaviors. Called meta-analysis, this approach, routinely used by biomedical researchers who want to draw robust conclusions about whether or not a particular therapy is effective, has been successfully adopted by behavioral ecologists. To note, meta-analysis is a statistical analysis of published results. Rather than simply saying that five species tolerate closer approaches when in larger groups while ten species flee sooner when in larger groups, a meta-analysis estimates the average effect size—the effect one variable has on another variable of interest—using all the evidence. It places more value in studies with more data and less value in studies with fewer data. By focusing on effect size, we understand the importance of a particular variable to explain the differences in an outcome—in our case, flight initiation distance.
Effect size, unlike traditional statistical significance, is not greatly influenced by the number of observations or amount of data you have. Rather, it’s describing the standardized consequence of a variable or a treatment. Contrast the effect on your longevity of smoking one cigarette or putting a single bullet in your head (please don’t do this at home!). You will survive much longer after smoking one cigarette because there is less uncertainty in the outcome of a bullet than a cigarette. This is not to say that heavy and regular smoking has no effect; it does. One estimate is that each cigarette takes eleven minutes off the lifespan of a heavy smoker. However, one bullet will most certainly ensure that you don’t make it to the next day. The effect size on longevity of a bullet is much greater than a single cigarette.
Meta-analyses can allow us to estimate the relative importance of different things that might explain differences in how soon species flee. A meta-analysis will tell us how important group size or sex or brain size or eye size is in explaining the differences we see in flight initiation distance. Meta-analyses will, therefore, allow us to identify the generally important costs and benefits of flight.
Ron and Larry focused on the highly dynamic decisions made by individuals, not species. But there are also differences among species, especially regarding the distance of a potential predator before flight. For instance, it’s much easier to walk up to a hummingbird than to an eagle. Why? What could explain these species-level differences? This was why Janice and I were walking toward different species of birds around the Tablelands. We were collecting a sufficiently large data set on FID from different species with very different life-history traits, including body size, typical group size, age at first reproduction, brain size, and longevity. All of these traits are likely to influence the ways that species allocate limited energy to survival, growth, and reproduction.
An insight gained from studying many species is that it’s possible to describe individuals and species as having relatively fast or slow life histories. Do they reproduce early, have relatively more offspring, and die young? Or do they mature slowly, delay the onset of first reproduction, have relatively few offspring, and live relatively long lives? Decisions animals make about allocating energy determine these life histories. An animal taking risks by allocating energy to rapid early growth and reproduction has a higher likelihood of death. By contrast, a cautious individual will make decisions that might slow its growth but increase its safety and survival rate. These cautious individuals will allocate more energy to each of their offspring, resulting in enhanced survival.
In nature we see species and individuals adopting both strategies because, ultimately, both can be effective ways to leave descendants. It depends on the specific environmental risks and on an individual’s future prospects. Individuals reared in dangerous environments poor in resources may ultimately do better by living short, risky lives and reproducing at a young age. Even if they are cautious in such an environment, they may not survive to live a long life.
My aim with our bird studies was to see how life-history traits may explain the differences in FID and thus how life history influences risk taking. I also included a number of natural-history traits, like habitat openness, to assess how the environment influences risk taking. Some habitats, like those of an Antarctic penguin, have good visibility, but others do not, like the dense forest where vireos live.
For my first comparative study I created a data set of 150 bird species with ten or more FID estimates for each species. I found that the initial distance of a human who walked toward an animal was the most important predictor of the distance the animal fled. If we began walking toward them from farther away, the birds were alerted but did not flee immediately.
This finding led me to propose the “flush early and avoid the rush” hypothesis—conveniently abbreviated FEAR. I hypothesized that animals that had to allocate their limited attention on monitoring our approach would begin to pay costs for this behavior at some point, and would benefit from moving away to reduce these monitoring costs. In other words, if animals were less able to eat, court mates, or monitor other potentially more important predators while keeping an eye on us, then they were paying a cost by monitoring us. Ultimately, those that fled early would not pay these costs and would allocate their time more efficiently. While not all species flee soon after detection, FEAR has been supported by both comparative analyses and meta-analyses. This means that FEAR is a generally important factor that influences escape decisions. Further, life-history and natural-history traits account for the differences in how close you can approach an animal before it flees.
After starting distance, the next most important variable explaining differences in FID is body size. Big birds, mammals, and lizards studied in places with few humans initiate escape sooner than smaller animals. But it is these very same large-bodied species that seem to habituate to humans if they are present and appear benign. We defined tolerance as the difference in FID at a place with many people and a place with fewer people. Thus, a tolerant species was one that let people approach much more closely in the area with many benign humans. The most important factor for tolerance of humans is the type of human activity in the study zone. We compared birds in urban areas and suburban areas, and those living inside and outside protected reserves, among other habitats. When animals were able to tolerate humans at all, the urban-rural contrast has the biggest effect; urbanization seems to make species more tolerant. Interestingly, the second most important variable explaining tolerance is body size. Large species that tolerate humans are the most adaptable.
Another factor that influences FID is predation risk. Birds who live in areas with more predators flee humans at greater distances. FID also decreases with sociality. More social species tolerate closer approaches, possibly because they live in larger groups. Additionally, body condition plays an important role in animals’ escape decisions. Parasitized birds tolerate closer approaches, possibly because it is more costly for them to flee. Hunted birds, wisely, flee at greater distances.
Brain size also has a profound effect on escape behavior in birds: birds with larger brains tolerate closer approaches. While larger species of birds have larger brains, we accounted for this statistically and suspect that larger-brained birds can tolerate closer approaches because they are better able to assess and evaluate risks. Once a risk is detected, animals must allocate their limited attention to monitoring the approaching threat. If this monitoring is too costly, animals should flee to reduce the ongoing costs of monitoring, the basis of the FEAR hypothesis. But the brain is the organ that assesses risk and allocates attention, and larger-brained species have the cognitive abilities to monitor risk while engaged in other activities.
Geographic location or environment is an important variable. A number of studies have shown that birds living in the tropics have a greater risk of dying young from predation due to the greater number and variety of predators in comparison to higher latitudes. This parallels other latitudinal trends showing many more species in the tropics than at the poles. The greater predation risk has selected for a whole suite of life-history responses—tropical birds typically live faster, invest less in more offspring, but die sooner than species living in temperate regions.
A fascinating finding is that you can get closer to female but not male birds at higher latitudes. Males maintained wariness across latitudes while females were less wary at higher latitudes. This was not explained by differences in body size between males and females, nor was it explained by differences in their color (many male birds are more brightly colored than females). What could explain this geographic and sex-specific result?
We interpreted this data as a result of reduced predation risk for nestlings at higher latitudes. Liana Zanette and Michael Clinchy’s results, presented in Chapter 3, were instrumental in this interpretation. Recall that birds hearing predator vocalizations reduced the number of feeding trips to their nests. If this is a generalizable idea, it makes concrete predictions about how risk should vary on an elevational gradient because altitude leads to the same patterns of predator diversity as latitude; there are fewer predators at high elevations and at higher latitudes. Thus, we may also expect fear, as measured by FID, to vary along an elevational gradient as well.
Lizard escape behavior is influenced by similar factors. Things that influence the cost of fleeing—such as food availability and social interactions—are the most important factors explaining differences in lizard escape decisions. Predator density is very important as well. We know that lizards and birds living with more predators are warier. As we expected from our understanding of avian fear, habitat factors also influence escape decisions, including how far lizards are from a refuge and whether or not they are in dense cover. Lizards far from their refuges escape at greater distances and those in dense cover remain there. Finally, the predator’s behavior also influences escape. Predators moving quickly toward prey drive prey to flee sooner.
Can the insights from these studies help address my initial question about why there is variation in species’ tolerance of people? As discussed, insights from FID studies have suggested some generalizations about the effects of body size. Large-bodied species in locations where they do not interact much with humans are more likely to be disturbed. Colleagues and I created a computer model that assumed that these distractions reduced the amount of time species could allocate to feeding and then looked at expected survival and reproduction. Disturbed animals are more likely to have reduced fitness when interrupted or agitated. However, this fitness cost drives selection for tolerance! Thus, large-bodied animals that are able to co-exist with people become more tolerant than smaller-bodied animals.
I wrote a draft of this chapter while on sabbatical in Sydney, Australia, where sacred ibis have become pest birds, better known as “bin chickens” in urban Sydney. Ibis stand almost two feet tall and have a long, curved bill. My favorite example of just how pesky they have become is an observation Janice made at Circular Quay in downtown Sydney. She sat watching a nearby ibis, perched behind a man reading his newspaper and eating his sandwich on a park bench. The ibis slowly inched its beak forward, around the man’s shoulder, and with a swift, arching movement grabbed the sandwich out of the man’s hand. Big birds. Big problems.
But what about the smaller birds? Do birds that become more tolerant of humans become more vulnerable to predators? My colleagues Diogo Samia, Benjamin Geffroy, Eduardo Bessa, and I tried to answer this question. In many cases predators avoid urban areas, leading to a “human shield” that protects urban prey from predation.
Human shields have been demonstrated to drive ecological and behavioral effects. They may trigger a cascade of events: predators avoid certain areas, and prey become more likely to frequent these areas. Prey then become less vigilant there because fewer predators hunt them. The reduction in vigilance means that prey can allocate more time to foraging, and the vegetation takes a noticeable hit. (We’ll revisit this idea in Chapter 9.) But both human shields and increased exposure to humans may be present in other situations. Nature-based tourism is one notable example. According to a recent report, over 8 billion people visit terrestrial protected areas annually. It’s as if each person on Earth visited a protected area once, and then some. And we know that there are deleterious consequences of this visitation: increased traffic and pollution, vegetation trampling, vehicular collisions with wildlife, and so on. Relatively small increases in mortality can tip populations from being seemingly stable to declining toward extinction. It would be ironic if the very eco- and nature-based tourism that seeks to protect nature is actually harming it.
I have always been a strong proponent of ecotourism. While studying marmots in Pakistan I wrote An Ecotourist’s Guide to Khunjerab National Park, a guidebook about the spectacularly beautiful high alpine national park in which I worked. I wrote it to help communicate with visitors and protect the park’s natural resources. Well-managed ecotourism tries to minimize the ecological impact of visitors while maximizing their positive effects: cultural preservation, but also resources and income for those who protect nature. Because ecotourists should want to minimize any negative impacts, identifying them is essential.
If, as my colleagues and I suggested, too many nature-based tourists made wildlife more vulnerable to predators, it would strengthen the argument against ecotourism. We hypothesized that tourists created a human shield that protected animals, leading them to become more docile and less responsive to human presence. The human shield effect has been implicated in increasing vulnerability to wildlife poachers and illegal hunters. But would it also, under the right circumstances, increase vulnerability to natural predators?
At first blush the idea seemed unlikely. We know that many prey have sophisticated contextual cues—habitat, weather, or location—to assess predation risk and modify their behavior to reduce exposure to predators. In some cases, prey initially retain abilities to recognize “past ghosts” when colonizing predator-free areas, but these abilities may be lost rapidly if there are indeed no predators (as discussed in Chapter 2). In addition, we know that many prey species have the ability to discriminate among different predator species and react according to the degree of risk they face from that type of predator. Thus, the idea that by habituating to people, prey may not recognize another species of predator appears unlikely.
Nevertheless, the idea that individuals who become excessively bold around humans may also be bold around their predators has some empirical support, and we expect selection to package similar sorts of traits together. For instance, docile individuals may be less responsive to both conspecifics and predators. Thus, if by being around humans animals are generally becoming more docile, which has been reported in some species, like samango monkeys, then perhaps we are creating animals less able to deal with their natural predators.
So what have we learned about economic logic? Quantifying the costs and benefits of behavior helps us ask whether or not behaviors are adaptive. We have seen that a variety of factors influence the distance at which an individual flees an approaching threat, and we have also identified patterns in this flight-initiation distance that are species specific. FID is quite sensitive to costs, and we have learned that the costs of escape have profound impacts on the distances at which animals flee. If there are costs to leaving a good foraging patch, a preferred perch, or a promising social interaction, animals can be approached more closely; they tolerate more risk.
Because fear is the side-effect of an economic decision, benefits also influence the risks we will accept. As an example, my son and I like to watch big wave surf competitions. Neither of us want to surf sixty-foot monster waves. When they crash, these waves create earth tremors that can be measured on the Richter scale. The super athletes who choose to surf them train by carrying boulders and weights underwater while holding their breath. They must prepare for possibly being held underwater and dragged across the ocean floor for over a minute. For this risk, the successful surfers win substantial prize money. In 2017, the top award for the Mavericks Challenge Big Wave competition in California was $120,000. In Teahupoo, Tahiti, those who survive the very large (but not truly enormous) waves breaking on an extremely shallow coral reef can win over half a million dollars. There are benefits beyond the initial financial reward. Winners and others who have photogenic rides decorate the covers of the world’s surf magazines and get lucrative sponsorship and advertising contracts. Thus, the benefits of successfully confronting fear in this scenario are obvious: great financial rewards. Would people stop surfing huge waves if the financial rewards were smaller? Probably not. After all, there is glory associated with mastering a huge wave and, at a very immediate level, the addiction of the rush.
Other decisions about accepting risk illustrate the economic arithmetic that we engage in when making decisions. I ask my students if they would go to the most dangerous place in the world, such as a war zone, if I paid them $100. Nobody accepts. However, as I up the ante they begin to squirm in their seats. $1000? $10,000? $100,000? $1,000,000? $10,000,000? For $10,000,000 they could afford a security detail, body armor, and other forms of protection. My students consider both the benefit of successfully confronting fears and the opportunity cost of not confronting their fears. We all have our price.
By understanding the economics of fear we gain insights into the trade-offs that both animals and humans make on a daily basis. These insights challenge us to seek strategies to successfully navigate them. But they also shed light on the effects of certain threatening environmental conditions in human society.
In humans, the logic of life-history theory has profound public health implications. If life itself is uncertain and resources are scarce, it makes evolutionary sense to reproduce early and often and invest fewer resources in each child. But this means the children themselves are at great risk. Those living in poverty would leave no descendants if they delayed reproducing too long, however, and died before they could do so. If a fetus or infant is stressed or experiences poor nutrition, the effects are seen throughout their lives. Research has shown that later in life, the pathologies of poverty include diabetes and heart disease, both of which are associated with reduced longevity. Ample resources and a stable environment provide the building blocks for maximal health in individuals and their descendants. We are thus descended from individuals who made the right decisions—those that maximized reproductive fitness—whether consciously or not.
Ultimately, however, we manage our risks and benefits, and this management is based on our perceptions of risks and rewards. We often get it wrong. Evaluating our perceptions is important work: it contributes to how we assess threats but also what will be needed to overcome irrational fears. US politicians who intend to instill fear by focusing on the rate of crimes committed by illegal immigrants often conveniently forget to note the higher rate of crimes committed by US citizens. Once people have made the association, however, it requires work to change their assessments to match reality.
Public perceptions of risk can also be influenced by a mistaken emphasis on a true but less relevant fact. Conservation biologists initially noted that the lionfish—accidentally introduced from the aquarium trade to the Caribbean—was poisonous, in addition to being highly invasive and destructive. This has inadvertently created problems for a creative and novel program to eliminate them—encouraging people to eat them! Potential consumers are thus wary of hunting or eating this really tasty fish, even though the risks of poisoning are eliminated if you handle them carefully with gloves. More work will be needed to overcome this manufactured fear.
Like the birds we’ve discussed in this chapter, we use economic logic to assess our risks, whether or not we are aware we are doing so. And for that we have to thank our ancestors. Their experiences prepared us to make the right decisions about how to maximize our benefits while reducing our costs in many situations. We are more cautious because of their experiences. A particular challenge arises when risks are truly novel. Consider identity theft—a truly novel risk with profound consequences for our security and well-being. We’ve not evolved the precise tools to be sufficiently skeptical of emails or phone calls that scare us into releasing our personal information. Novel threats, as we will learn in Chapter 9, require novel responses, and these may require some work. But for those who are anxious or fearful about threats, for the right price, these fears can be overcome. And this is good because we live in a world surrounded by risk.