Information

Is there any example, where cooperative behaviour of predators induce fear in prey population?

Is there any example, where cooperative behaviour of predators induce fear in prey population?


We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

I am basically from math background and doing Ph.D in mathematical biology. So I am not so efficient in biology. But my problem based on cooperative behaviour of animals during hunting which is related to the animal behaviours in ecology.

We know that different animals show cooperative behaviour during hunting (or, group hunting) such as wild dogs, lions, chimpanzees, birds, spiders, crocodiles etc. We also know that fear of predator (predation risk) reduces the growth rate of prey population. Now, I am looking for example(s), in which prey population does not fear predator but cooperative behaviour of predators induces (creates) fear in prey.


Fear of predators may be a bigger killer than the predators themselves

(PhysOrg.com) -- When biologists consider the effects that predators have on their prey, they shouldn’t just count the number of individuals consumed. According to a University of Rhode Island ecologist, they must also examine the effects of fear.

URI Assistant Professor Evan Preisser said that fear of being eaten can reduce population densities as much or even more than the actual quantities of individuals killed by predators.

“Prey are far from helpless victims of predators,” said Preisser. “They employ a wide array of defensive strategies to protect themselves. But the costs of these strategies may have a larger impact on their population than the direct effect of being eaten.”

To avoid being consumed by a predator, many prey species will spend more time hiding and less time eating. This can lead to a lower body mass, reduced reproduction rates, fewer offspring, and a lower rate of survival.

Preisser notes that fire ants, for example, are highly successful at finding resources, but they are “totally freaked out” by a species of parasitoid fly that lays its eggs inside the ants, which ultimately kills them.

“If one of these flies comes along, all the ants will hide and remain hidden for a really long time,” he said. “Research by Donald Feener at the University of Utah has shown that the flies actually have a very low success rate at killing the ants because the ants are so good at hiding. They spend so much time in hiding, however, that the whole ant population becomes weaker.”

Preisser also points to research conducted by Oswald Schmitz at Yale University documenting that grasshoppers can be so afraid of wolf spiders that they will starve to death rather than come out of hiding and feed in the presence of the spider.

In a research paper published in the journal PLoS ONE in June, Preisser and Daniel Bolnick of the University of Texas-Austin found that the presence of a predator reduces prey feeding rates and overall activity rates by 57 percent and 45 percent, respectively, among species living in aquatic ecosystems and by 45 percent and 34 percent among those in terrestrial ecosystems.

“Just the simple presence of a predator can increase the mortality of prey species by as much as five percent,” Preisser said.

The URI ecologist is the guest editor of an upcoming special, three-article feature on “nonconsumptive predator effects on prey dynamics” in the September issue of Ecology, the journal of the Ecological Society of America. The articles were written as a result of a working group convened by the National Center for Ecological Analysis and Synthesis to examine the topic, which Preisser and Bolnick co-chaired.

In his introduction to the feature, Preisser wrote that fear-based, nonconsumptive effectives of predation “may extend a predator’s reach far beyond its grasp.”


Introduction

Predators capture, kill, and consume their prey. This apparently trivial statement has complex and important implications: predators reduce prey population density, which in turn can affect the population growth of prey resources and other predators. The ‘cascading’ effects of consumption propagate throughout ecological communities, and are critically important to community dynamics. Food web models attempt to capture these dynamics using simultaneous population dynamic equations that link predators, prey, and resources via consumption rates [1]. Basic predator-prey models illustrate the primacy of consumption in structuring ecological thinking about food webs. For example, the classic Lotka-Volterra equations [2], [3] describe the interaction between predator (P) and prey (N) population densities as follows: (1) (2) where a is the capture rate, e is the rate at which offspring are produced per unit of energy income (aNP) into progeny, d is the predator death rate, and r is the prey intrinsic growth rate. In this model, predators and prey interact solely through successful predator attacks. Predators cannot grow without reducing prey density similarly, any reduction in prey populations must contribute to predator population growth.

Both theoretical and empirical studies have challenged ecology's focus on consumption (a) in predator-prey dynamics [reviewed in 4,5–8]. The mere threat of predation can be sufficient to reduce prey growth, survival, or fecundity [although the strength of these responses can certainly be affected by factors such as predator hunting mode 9,10]. Far from being passive players on the ecological stage, prey employ a suite of behavioral and morphological defenses to avoid predation. These defensive strategies often have significant costs that reduce prey fecundity or survival. Predators can thus affect prey populations both by direct consumption (consumptive effects, ‘CE’) and by inducing costly defensive changes in prey traits [nonconsumptive effects, ‘NCE’, also called ‘trait-mediated effects’ by some researchers 11].

In the past decade, numerous experimental studies have measured the magnitude of NCE, a previously underappreciated class of predator effects. The results have been quite variable: NCE appear to dominate some interactions [e.g., 12], and be weak or absent in others [e.g., 13]. Despite this heterogeneity, it has become clear that NCE are common and that they can have powerful effects on prey populations. A recent meta-analysis found that NCE often rival the effect of CE on prey populations [14]. This result aligns with theoretical predictions [8], [15]–[17] that NCE play an important role in structuring predator-prey interactions, and suggests that the exclusive focus on consumption characteristic of most predator-prey models (Lotka-Volterra and its offshoots) can be misleading.

The integration of NCE into predator-prey theory is clearly critical for progress towards a comprehensive understanding of predator-prey dynamics. We seek to assist such integration by distinguishing among multiple biologically distinct mechanisms by which predator intimidation can affect prey density and, thus, population dynamics. Our hope is that distinguishing among these mechanisms may make it easier to contrast outcomes from different study systems, draw general inferences, and more effectively connect empirical results with predictive models. We develop this framework by systematically expanding the term in the Lotka-Volterra equation through which the effects of predator NCE are manifest, the intrinsic growth rate (r). Our intent is not a mathematical analysis of the resulting framework rather, we examine the different ways in which r might respond to predator density, using the Lotka-Volterra model as a heuristic tool for identifying potential NCE pathways and clarifying the relationships between them. Because different types of prey defenses may have distinct ecological effects, genetic causes, and evolutionary consequences, such a classification scheme may help distinguish between biologically distinct types of interactions, allowing a more comprehensive and more mechanistic description of predator-prey dynamics. We integrate our explication of this framework with a metaanalysis of published literature assessing (where possible) the mean effect sizes of both the individual and combined pathways of nonconsumptive predator effects.


The Fear Factor: How the Peril of Predators Can Transform a Landscape

Biologists are developing a deeper understanding of how the terror felt by prey creates a major — and greatly under-appreciated — behavioral dynamic that ripples through all kinds of ecosystems.

On a beach in the Gulf Islands off the coast of southern British Columbia, biologist Liana Zanette lashed speakers to a tree for a special broadcast. It wasn’t music, and it wasn’t played for people. It was the sharp sound of barking dogs — aimed at a population of bold raccoons.

Instead of being nocturnal, as raccoons usually are, these animals had also grown accustomed to feeding during the day. Wandering away from the security of forests, they traveled far out onto the exposed tidal flats to look for worms, clams, and other food, unconcerned about being eaten because predators in the area are rare. Bears and wolves had long been extirpated. Only the occasional dog brought there by locals threatened their utopia.

So for two months, Zanette blasted out two sounds — barking dogs, and, as a control, barking seals and sea lions, which posed no predation threat.

The effort to restore a semblance of fear to this raccoon nirvana enjoyed marked success. As Zanette reported in a paper in Nature Communications, the raccoons spent two-thirds less time feeding in the tidal zones when the electronic dogs barked, which resulted in 81 percent more fish being found in the pools and some 60 percent more worms and red rock crabs. “It’s a massive effect,” says Zanette, a biologist at Western University in Ontario.

Welcome to the ecology of fear. Many biologists think that the terror felt by prey creates a major — and greatly under-appreciated — behavioral dynamic that ripples through all kinds of ecosystems. In fact, it likely has more impact on an ecosystem than the actual predation itself, because one snarling predator can change the feeding behavior of many individuals without killing them.

“All animals, no matter what taxa, need to worry about predators,” says Zanette. “Even tigers worry about humans. So it’s a really strong evolutionary force.”

Because it’s hard to separate from other elements, though, the fear factor has been difficult to study and so isn’t accounted for in ecological models. That’s changing.

The 1995 reintroduction of wolves to Yellowstone National Park led to seminal research on the ecology of fear, with the wolves altering the feeding behavior of elk and touching off a cascade of ecological effects on vegetation and other wildlife. In recent years, researchers like Zanette have developed new approaches in hopes of creating a deeper and more nuanced understanding of the role of fear, given how important it is to understanding ecosystems. Yellowstone, meanwhile, continues to provide new insights into the ecology of fear on a large, wild landscape.

Fear of wolves altered the feeding behavior of elk in Yellowstone, with important implications for the ecosystem.

The topic raises important conservation questions. As predators have declined globally, their disappearance — and the fear factor they induce in prey — has had knock-on effects in numerous ecosystems. Can a decline in biodiversity brought about by a lack of predators be restored by bringing them back?

“It underscores the importance of protecting intact populations of predators,” says Larry Dill, an emeritus professor of biology at Simon Fraser University who studied similar dynamics in marine ecosystems. “Without healthy populations of wolves and sharks you have impacts all the way down the food chain.”

Among the first studies of the role of fear were small-scale efforts in controlled environments with spiders and grasshoppers in the 1980s. Oswald Schmitz of the Yale School of Forestry & Environmental Studies watched as grasshoppers foraged inside screen cages, both with and without predatory spiders present. When the spiders were absent, he noticed the hoppers foraged on grass when the spiders were returned to the cages, the grasshoppers switched to feed on forbs — taller flowering plants that offered the grasshoppers both food and a place to hide.

In a later experiment, Schmitz created what he called “risk spiders.” These were predatory spiders, but with their mouth parts glued shut so they still appeared to present a risk to the grasshopper, but couldn’t actually prey on them. And even though the spiders had been de-fanged, the hoppers still fled to the forbs, showing, Schmitz concluded, that the mere presence of predators — and the fear they engender — affected the plant-eating behavior of the prey and changed ecosystem dynamics.

That, however, was on a very small scale in the lab. The ecology of fear gained prominence in the 2000s when Yellowstone researchers noticed that willows and aspen had regrown in some places following the return of wolves. The wolf, an apex predator, killed so many of the park’s elk, experts said, that it caused profound adjustments to the ecosystem — something known as a trophic cascade. But fear also significantly altered the feeding of the surviving elk, with important implications for the Yellowstone ecosystem.

The steep drop in elk populations — from roughly 20,000 before the wolf reintroduction to 6,000 today — meant that the ungulates were mowing down fewer of the park’s young willow and aspen trees. With willows, aspen, and grass more abundant, a range of other species, from beaver to songbirds, benefited from the expansion of their prime habitat. More beaver ponds meant higher water tables, more stabilized stream flows, and more and better fish habitat.

But some studies showed that the fear of predation also had a profound impact on the ecosystem. When the wolves showed up, the elk suddenly snapped to attention. They had to be extremely vigilant, ecologists said, and could no longer hang out carefree by the river eating riparian plants to their heart’s content.

That meant grazing intensity and patterns shifted. With elk no longer feeling secure as they browsed in the river bottoms, willow and aspen trees grew tall again for the first time in decades. One 2010 paper on this new psychological terrain of Yellowstone was titled “The Landscape of Fear: The Ecological Implications of Being Afraid.”

Given that the wolf is a charismatic animal and Yellowstone National Park is world renowned, the “landscape of fear” became famous. A YouTube video narrated by Guardian columnist George Monbiot, for example, entitled How Wolves Change Rivers, has more than 35 million views — and vastly overstates the case, many scientists say.

Some critics say the notion of wolves restoring Yellowstone by dint of the terror they engender is a great story that has captured people’s imagination. But critics such as Oswald Schmitz are not convinced that these changes are linked to the reintroduction of wolves and the subsequent behavioral changes of elk. The growth of more willow and aspen, for example, could be accounted for by more flooding, which means more water for the streamside vegetation. But that possibility, some critics have said, hasn’t been factored into the studies. And Schmitz says there is good evidence that elk are not all that worried about the presence of wolves.

The debate over the ecology of fear ‘isn’t whether it’s true, but how it works,’ says one researcher.

Doug Smith, who has been the wolf biologist in Yellowstone park since the animals were brought back, disagrees with critics who dismiss the ecological impact created by the elks’ fear of wolves. While the fear effect “has been popularized beyond its real impact,” he told me when I visited the park recently, “there is a big pearl of truth to the trophic cascade” caused by the wolves’ presence.

In his opinion, “the debate now isn’t over whether it’s true, but how it works.” Disentangling factors such as fear from other impacts, such as bottomland flooding or the greatly reduced numbers of elk, is difficult, he says. But in a paper on the subject due out later this year, in which researchers tracked GPS-collared elk near wolves, he and colleagues found that “elk avoid risky habitat in willow stands during peak wolf activity. Fear of wolves shapes the landscape.” But precisely how is “a lot more complicated than we think.”

Such research could provide support for the protection of existing predator populations or a rationale to introduce new ones. Understanding the fear landscape may also offer new ways to manage ecosystems. Biologists conserving sage grouse in Oregon, for example, have found the ground-nesting birds avoid settling near juniper forests because ravens and raptors perch there and prey on the grouse. In many sage grouse nest sites, thousands of junipers have been cut down.

In the meantime, don’t look for predators to restore the broader natural world. The impact of wolves on a wild landscape like Yellowstone is one thing, but landscapes significantly changed by humans represent a much different question.

Meanwhile, another question beginning to be asked by researchers is what effect does the fear that humans engender in wildlife have on the natural world? A 2015 study called humans “super predators” because they kill carnivores at a whopping nine times the rate of other predators. Attuned to the steep risk posed by people, how do wolves, bears, wolverines, and bobcats change their behavior as a result of human presence, and how does that alter their role in the ecosystem?

No one knows, but Zanette has weighed in on this question as well. In the UK, badgers — which love worms and sometimes prey on birds, animals, and reptiles — have had only one primary predator for centuries: humans. Zanette and her colleagues set out buckets of peanuts mixed with soil for the badgers. The scientists then played recordings first of sheep, and then of wolves, which are long gone from the ecosystem. Badgers ignored both in their feeding. Recordings of bears and dogs generated some delay in heading to the buckets. “But the thing they were most afraid of was people talking,” Zanette says. “We played BBC broadcasts, documentaries on aviation, actors reading books, and they generally did not come out of their hole while the humans were talking.”

Schmitz says the impacts of the presence of people — and the landscape of fear they create — could be substantial. “Humans,” he says, “can have important effects by scaring predators away, to the extent that ecosystems’ processes could fundamentally change.”


Predators indirectly induce stronger prey through a trophic cascade

Many prey species induce defences in direct response to predation cues. However, prey defences could also be enhanced by predators indirectly via mechanisms that increase resource availability to prey, e.g. trophic cascades. We evaluated the relative impacts of these direct and indirect effects on the mechanical strength of the New Zealand sea urchin Evechinus chloroticus. We measured crush-resistance of sea urchin tests (skeletons) in (i) two marine reserves, where predators of sea urchins are relatively common and have initiated a trophic cascade resulting in abundant food for surviving urchins in the form of kelp, and (ii) two adjacent fished areas where predators and kelps are rare. Sea urchins inhabiting protected rocky reefs with abundant predators and food had more crush-resistant tests than individuals on nearby fished reefs where predators and food were relatively rare. A six-month long mesocosm experiment showed that while both food supply and predator cues increased crush-resistance, the positive effect of food supply on crush-resistance was greater. Our results demonstrate a novel mechanism whereby a putative morphological defence in a prey species is indirectly strengthened by predators via cascading predator effects on resource availability. This potentially represents an important mechanism that promotes prey persistence in the presence of predators.

1. Introduction

Many organisms actively induce morphological defences in response to predators [1]. Such morphological changes are widely considered to be a direct response to the detection of mechanical, tactile, visual or chemical cues from predators. However, it has been suggested that in many cases induced defences occur as a passive by-product of changes in prey behaviour or resource availability, rather than being a direct morphological response to predation cues [2]. Given that predators can strongly influence prey behaviour [3], and also alter resource availability to prey through trophic cascades [4], there is plenty of scope for morphological defences in prey to be induced by predators via indirect mechanisms.

Sea urchins are important herbivores in many shallow benthic habitats [5] and are often implicated in trophic cascades [6]. They are encased in a test (skeleton) that is porous, but remarkably strong [7], and confers at least some protection against predators that crush their prey [8]. Calcite spines project from the test surface, helping to repel predators [9], spread impact loads [10] and capture food [11]. These features are highly plastic and can differ morphologically depending on variation in abiotic factors such as water motion [12], or biotic factors such as food availability [13,14] and predation cues [15]. Predators can induce morphological defences in sea urchins via unsuccessful attacks or chemical cues released by predators or injured prey. For example, waterborne cues from predatory crabs induce thicker tests in Strongylocentrotus droebachiensis [15]. Within a marine reserve in northern New Zealand, where predators were abundant, Cole & Keuskamp [16] found that sea urchins had heavier tests than in adjacent fished areas, where predators were rare. While they did not test the mechanism responsible they suggested that this was either due to increased calcification in response to sublethal attacks by predators or selective predation on sea urchins with thin tests.

Predators also have the potential to induce these defensive features in sea urchins indirectly through their effects on sea urchin behaviour [17] and cascading impacts on the availability of kelp, a major food for many sea urchins [18]. When predators are rare (often due to overfishing), sea urchins can reach high densities and create ‘urchin barrens’, in which kelp and other food is scarce [19]. Starving sea urchins invest in their feeding apparatus at the expense of other body parts, including the test [20,21]. Some sea urchins even shrink when food-limited, as they reallocate calcite from the test to their jaws [22]. When predators are sufficiently large and abundant they initiate a trophic cascade, whereby sea urchins are either eaten or restricted to crevices where they eat little living kelp, allowing the kelp to recover [23]. In food-rich habitats sea urchins tend to have thicker tests [24,25], which potentially help the urchins persist in the face of strong predator pressure. It is therefore possible that sea urchins may develop thicker tests as an indirect response to predators mediated through a trophic cascade that increases food availability to sea urchins.

Our study aimed to examine the relative importance of predator cues and food availability on the development of putative morphological defences in sea urchins. We compared morphological attributes of sea urchins inside and outside two well-established marine reserves in north-eastern New Zealand. Previous research has demonstrated that sea urchin predators are larger and more abundant in these reserves compared to adjacent fished areas, and this results in lower sea urchin and increased kelp densities via a trophic cascade [23,26]. This therefore provides a large-scale experimental framework to investigate the effects of predators on sea urchin morphology. To determine whether differences in morphological defences observed between reserve and fished areas were a direct response to the presence of predation cues or an indirect response to food availability a six-month long mesocosm experiment, where sea urchins were subjected to different levels of predation cues and food availability, was carried out.

2. Material and methods

(a) Morphological variation in sea urchins

Morphological attributes of the endemic sea urchin Evechinus chloroticus were quantified in two marine reserves in north-eastern New Zealand, and in adjacent fished areas with similar topographies and exposures to wave action ([27] see electronic supplementary material, figure S1 for locations). The Cape Rodney-Okakari Point or Leigh Marine Reserve, established in 1975, is 549 ha, while the Tawharanui Marine Reserve, established in 1982, is 350 ha. All marine life is fully protected inside these reserves and abundances of the key sea urchin predators snapper Pagrus auratus and red rock lobster Jasus edwardsii are considerably higher than on adjacent reefs [27] where they are harvested commercially and recreationally. Evechinus chloroticus is lightly harvested from the shallow subtidal reefs recreationally. Dense forests of kelp Ecklonia radiata occur inside both marine reserves at depths of 4–6 m, while similar depths on fished reefs are virtually devoid of Ecklonia radiata and are characterized by urchin barrens [17,28]. The lower abundances of sea urchins and higher abundances of kelp inside these marine reserves relative to adjacent fished areas has been attributed to the cascading effects of predatory fish and lobsters [23,26].

During June and July 2013 Evechinus chloroticus of a range of sizes were collected (n = 20) from 4–6 m depth at four sites inside and outside each of the Leigh and Tawharanui marine reserves (see electronic supplementary material, figure S1). Collection of sea urchins from marine reserves was permitted by the New Zealand Department of Conservation, under research grant 4560. All Evechinus chloroticus were held in flow-through seawater tanks at the University of Auckland's Leigh Marine Laboratory for a maximum of 18 h before processing. Test diameter was measured using callipers (±1 mm). Four primary spines were removed from the equator of the test and their lengths measured (±0.1 mm). Each sea urchin was placed on a device designed to test crush-resistance, a measure of strength considered relevant to predation by snapper, which take an entire sea urchin in their mouth and bite down until the test is crushed (N.T.S. 2004, personal observation). The device consisted of an analogue set of scales, calibrated using a known weight, sitting underneath a bracket with a metal shaft and foot (maximum surface contact area 30 cm 2 ) (electronic supplementary material, figure S2). The shaft was wound down onto the aboral surface of a sea urchin at a constant speed until the test cracked, evident by a loud pop and sudden release in pressure. A slider indicated the load (±1.0 kg) required to crack the test. Spines absorb the impact of static loading [10] and also caused the foot to slip from the centre of the aboral surface, resulting in pressure being applied heavily to one side of the sea urchin only. To counter this, all spines on the aboral surface were trimmed back to the test using scissors before crushing. The thickness of the cracked test was taken as the average of four measurements made at randomly chosen points using callipers (±0.1 mm).

(b) Induction of putative defences

To determine whether food availability and/or predation cues were responsible for the morphological variation observed in the field survey, a mesocosm experiment was run using juvenile Evechinus chloroticus over six months from June to December 2013. In May 2013 juvenile sea urchins (less than 40 mm test diameter) were collected from a fished reef (Nordic Reef, electronic supplementary material, figure S1, 36°17′35.19″ S, 174°48′35.54″ E) and transported back to the Leigh Marine Laboratory fully submerged. We used juvenile urchins collected from urchin barren habitat at a fished reef, as individuals from a marine reserve could have already developed crush-resistance due to prior exposure to predation cues or enhanced food. Juvenile sea urchins were held in a flow-through 1500-l tank with small cobbles covered in crustose coralline algae on the bottom of the tank. This replicated an urchin barren habitat where urchins have limited access to food but can survive by grazing coralline algae and microalgae growing on the rocky substratum. Urchins were held in this tank for four weeks, without any additional feeding, so that they all started the experiment at a similar level of hunger.

The sea urchins used for the experiment (n = 256) had initial test diameters of 16–29 mm (±1 mm). Individuals were ordered by size, and starting with the smallest, a single individual was added to each of 32 replicate buckets, then another individual to each of the 32 buckets, until each bucket contained eight individuals. The individual tanks were 10-l buckets, each containing a small coralline algae-covered rock to provide habitat for the juvenile urchins. The predation cue used in this experiment was a crushed conspecific, which elicits a strong behavioural response in E. chloroticus [17] and in other sea urchin species [29–32]. To stop the sea urchins climbing up the sides of the bucket, while allowing for the addition of predation cues, a clear plastic lid with holes drilled in it was positioned 95 mm from the bottom of each bucket. Seawater was supplied via a 10-mm pipe passing through the lid. A short length of narrow tubing connected to the side of the pipe just above the lid drew water containing the predation cue down into the bucket via the venturi effect. The experiment was run in outdoor tanks in a continuous flow seawater system and under 80% shade cloth to mimic natural light levels in the field.

Sea urchins were subjected to two treatments (food and predation cue), each with two levels in a fully-crossed design. Each group had eight replicate buckets with eight sea urchins in each. The ‘food’ treatment had two levels which aimed to represent the different food availability between a kelp forest (‘+food’), where drift kelp is readily available, and an urchin barren (‘−food’), where kelp is absent but may occasionally occur as drift. In both cases sea urchins also had access to crustose coralline algae and microalgae on the small rock in each bucket. Juvenile sea urchins in the ‘+food’ treatment were given one blade of fresh Ecklonia radiata weekly, while those in the ‘−food’ treatment were given one blade per month. Uneaten kelp fragments remaining in containers at the next feeding date were left there to be eaten. We noted that there was rarely any uneaten kelp remaining in the ‘low-food’ replicates, and while there was some uneaten kelp in about half the ‘high food’ replicates it did not accumulate over time (A.J.P.S. 2013, personal observation). Since the experiment ran for six months, we consider that uneaten food at the end of the experiment was a negligible proportion of the total amount provided, so it was reasonable to assume that individuals in the high-food replicates ate approximately four times more food than those in the low-food replicates (ignoring the unquantifiable amount of food the urchins may have obtained from algae growing on the small rocks that were provided for shelter in each container). Each E. radiata blade was ca 200 × 40 mm and had a blotted mass of ca 4.3 g, so each of the eight individual urchins per replicate was provided with 0.08 g kelp d −1 in the ‘+food’ treatment and 0.02 g kelp d −1 in the ‘−food’ treatment. In a pilot feeding assay, run for 4 days during the previous summer when seawater temperatures were about 20°C, 20-mm urchins ate an average of 0.07 g kelp d −1 and 30-mm urchins ate 0.15 g kelp d −1 when provided with surplus kelp. These feeding rates would likely have been lower over the course of our experiment because water temperatures were lower than they were in the feeding assay (down to 14°C), so we consider that the urchins in the ‘+food’ treatment received adequate food.

For the predation cue treatment, a large E. chloroticus was cracked in half, and one half was placed on the lid of a bucket (‘+predation cue’). This was done twice a week with the cracked half remaining until being replaced by a new one. Control buckets were not subjected to crushed conspecifics (‘−predation cue’). We assumed that E. chloroticus would detect the crushed conspecifics given the strong behavioural response to the same cue in a previous mesocosm experiment that used much larger tanks (1500 l versus 10 l in the present experiment), where the cue would have been far more dilute [17].

At the completion of the experiment sea urchins were processed to obtain size, test thickness, resistance to crushing and spine length, using the same methods described above except that a finer-scale crushing device was employed due to the small size of the sea urchins. This was a hollow piston (weighing 2.2 kg) that could be filled with water. Water was slowly added until the sea urchin, placed directly underneath the piston foot, was crushed. The water was weighed to gain the total (including the piston weight) crush-resistance (±0.1 kg). The largest individuals required one or two 1-kg lead weights to be added to the piston prior to crushing, as not enough water could be held inside the piston.

An environmental scanning electron microscope (eSEM) was used to examine differences in the microstructure of the sea urchin test between treatments. The five largest sea urchins (approx. 30 mm) were selected from each of the four treatment groups. These were dried for 48 h at a constant temperature of 60°C, then soaked for approximately 48 h in NaOH to remove any organic material [33]. Sea urchins were soaked in a 3 M NaOH solution for 18 h, removed then re-soaked in weaker 2M NaOH solution for a further 24 h. An interambulacral plate was then removed from each sea urchin. A small section from this plate's equator was broken off and mounted on a slide then platinum-coated prior to examination under a Quanta eSEM. The undamaged face of each piece of test was examined under SEM, not the fracture line itself.

A series of images were taken across the face of the interambulacral plate and analysed using the particle analysis tool in ImageJ v. 1.44. Plate structure varied between the edge and centre of the plate, therefore any images containing edges were excluded from the analysis. One hundred non-overlapping quadrats (100 × 100 µm = 0.01 mm 2 ) were equally divided between the remaining images. Within each, the number of pores, the average pore size (±0.1 µm diameter) and the total pore space were measured using the particle analysis tool.

(c) Data analysis

All statistical analyses were carried out using R v. 3.0.2. To analyse sea urchin morphological attributes linear mixed models were fitted with random slopes and intercepts using REML. Test diameter nested within Site was a random factor, and Reserve (yes or no), Location (Leigh or Tawharanui) and Test diameter were fixed factors. Homogeneity of variance and normal distribution of each variable were visually assessed prior to analysis by plotting residuals against fitted values. We used an information-theoretic model selection approach and model averaging to determine the relative importance of main effects and interaction terms in explaining variation in crush-resistance, test thickness and spine length. We conducted ‘all possible’ combination model selection based on small sample corrected Akaike's information criterion (AICc) using the ‘dredge’ function in the R package MuMIn. All models within two AICc units of the best-fit model [34] were used to estimate the final model coefficients using model averaging [35].

To analyse morphological attributes of juvenile sea urchins in the mesocosm experiment, linear mixed models were fitted with random slopes using REML. Bucket was treated as a random factor and food (+ or −), predator (+ or −) and test diameter were fixed factors. The same model selection approach as above was used to identify the main determinants of the morphological attributes of juvenile sea urchins.

Differences in growth and test microstructure of juvenile sea urchins were analysed using two-way ANOVA. The factors for this were food (+ or −) and predation cue (+ or −).

3. Results

(a) Morphological variation in sea urchins

All of the best-supported models for crush-resistance included Location, Reserve and Test diameter (table 1). Crush-resistance of sea urchins increased with test diameter, and was higher in individuals from marine reserves than those from fished reefs by an average of 13.4 ± 5.0 (s.e.) kg (figure 1a,b and table 1). There was also some support for an important interaction between Location and Reserve (table 1, included in four models and RVI = 0.62), and the effect of Reserve on test strength was greater at Leigh (16.4 ± 5.3 kg t = 3.1, p = 0.021) than Tawharanui (7.0 ± 2.1 kg t = 3.10, p = 0.017).

Figure 1. Crush-resistance (a,b), test thickness (c,d) and primary spine length (e,f) for the sea urchin Evechinus chloroticus from marine reserves and fished reefs at Leigh and Tawharanui, north-eastern New Zealand.

Table 1. Model-averaged coefficients for linear mixed models of the effect of Reserve (yes or no), Location (Leigh or Tawharanui) and test diameter on crush-resistance, test thickness and spine length in the sea urchin Evechinus chloroticus. Coefficients are averaged by model weight across the best-supported models, i.e. within AICc < 2 of the best model. ‘RVI’ is the relative variable importance scores, and ‘N’ indicates the number of best-supported models in which the term was included. Note that terms that were not included in the best supported models are not shown. Bold values indicate significant terms (α = 0.05).

Test thickness also increased with size (figure 1c,d). The best-supported models all included Location, Reserve, Size and Location*Reserve (table 1). Sea urchins inside the marine reserve at Leigh had thicker tests than similar-sized individuals on the surrounding fished reefs (by 0.23 ± 0.07 mm t = 3.3, p = 0.016), but there was no difference in test thickness between sea urchins inside and outside of the marine reserve at Tawharanui (t = 1.1, p = 0.305).

Spine length increased significantly with test diameter (figure 1e,f and table 1). Spine length did not vary significantly with reserve status or location.

(b) Induction of putative defences

The effect of food on juvenile sea urchin growth was not quite significant at the α = 0.05 level (F1,28 = 4.123, p = 0.052). In the +food treatment juvenile sea urchins grew 0.67 ± 0.05 mm month −1 , versus 0.52 ± 0.05 mm month −1 in the −food treatment. There was no effect of predator cues on juvenile growth (F1,28 = 1.849, p = 0.185) and no food*predator interaction (F1,28 = 0.198, p = 0.660) (electronic supplementary material, figure S3).

For crush-resistance both food and predation cues interacted significantly with test diameter and were included in all of the best-supported models (figure 2a and table 2). At large sizes urchins with food and with predator cues were more crush-resistant. Food had a greater effect than predation cues on overall crush-resistance as indicated by the 2.4 times larger slope coefficient for food*test diameter (figure 2a and table 2).

Figure 2. Crush-resistance (a), test thickness (b) and primary spine length (c) of sea urchins Evechinus chloroticus grown under different food and predation cue conditions in a mesocosm experiment. Solid regression lines indicate −predation cue treatments, dashed regression lines indicate +predation cue treatments.

Table 2. Model-averaged coefficients for linear mixed models of food (+/−) predation cue (+/−) and test diameter on crush-resistance, test thickness and spine length of juvenile sea urchins, Evechinus chloroticus, grown in a mesocosm experiment. Results shown are for the optimal model, therefore only significant interaction terms are shown for each morphological feature investigated. Coefficients are averaged by model weight across the best-supported models, i.e. within AICc < 2 of the best model. ‘RVI’ is the relative variable importance scores, and ‘N’ indicates the number of best-supported models in which the term was included. Note that terms that were not included in the best supported models are not shown. Bold values indicate significant terms (α = 0.05).

For test thickness the interaction between predation cue and test diameter was included in all of the best-supported models (figure 2b and table 2). When predator cues were present large urchins tended to have thinner tests than when there was no predation cue. Food did not have a significant effect on test thickness.

The best models for spine length all included interactions between test diameter and both food and predation cues (figure 2c and table 2). Larger sea urchins in both food and predation cue treatments tended to have longer spines than when predation cues or food were not present. The coefficient for the food*test diameter interaction was 1.8 times larger than the predator*test diameter.

Pore density did not differ significantly between the four treatment groups (figure 3a), regardless of food (F1,19 = 0.862 p = 0.376) or predation cues (F1,19 = 0.188, p = 0.680). Sea urchins fed weekly did, however, have significantly smaller pores (13.18 ± 0.32 µm 2 versus 14.91 ± 0.38 µm 2 , figures 3b and 4, F1,19 = 9.681, p = 0.007) and less total pore space (11.80 ± 0.44% versus 14.11 ± 0.36%, figures 3c and 4, F1,19 = 14.004, p < 0.002) than those fed monthly. Predation cues did not influence either feature (F1,19 = 0.000, p = 0.992 and F1,19 = 0.240, p < 0.631, respectively).

Figure 3. Number of pores (a), pore size (b) and total pore space (c) in the interambulacral plate of sea urchins Evechinus chloroticus grown under different food and predation cue conditions in a mesocosm experiment. Bars represent average +1 s.e.

Figure 4. eSEM images of the porous surface of an interambulacral plate from sea urchins grown under different food and predation cue conditions in a mesocosm experiment. Sea urchins fed weekly (+) food had smaller pores than those fed monthly (−).

At the level of individual sea urchins, crush-resistance was negatively related to total pore space, while there was no significant relationship between crush-resistance and test thickness (electronic supplementary material, figure S4).

4. Discussion

We found that sea urchins inside marine reserves where predators and kelp were abundant had more crush-resistant tests than similar-sized individuals from adjacent fished coastlines where predators and kelp were rare. Results from the mesocosm experiment demonstrated that food availability was the main determinant of crush-resistance, with predation cues having a weaker effect. This indicates that the greater crush-resistance of sea urchins in marine reserves is not primarily a direct response to predation as previously suggested [16], but instead is mostly an indirect effect whereby the presence of predators leads to greater food availability through a trophic cascade. In the marine reserves examined in this study a trophic cascade has been clearly demonstrated [23,26], and recent research has further demonstrated that behavioural changes of sea urchins (to a more crevice-dwelling lifestyle) in response to predators contribute to this trophic cascade [17]. While a reduction in feeding rates due to a ‘fear effect’ on foraging behaviour can to lead to passive development of morphological defences in some prey species [36,37], this mechanism is unlikely to generate stronger tests in sea urchins, which typically weaken when food is limiting ([14,24,25] this study). Furthermore, while sea urchins are more likely to be restricted to crevices in the presence of predators, kelp was highly abundant on the surrounding reefs [17] and sea urchins have frequent access to drifting kelp fragments [38]. To our knowledge, this is the first example of mechanical strength being induced indirectly by predators through a trophic pathway that increases overall food availability. (We follow Bourdeau [37] in using the term ‘induction’ to describe both direct and indirect routes through which predators strengthen putative prey defences.)

In our study, all sizes of sea urchins inside marine reserves had more crush-resistant tests than comparably-sized individuals on fished reefs. Predatory fishes often attack sea urchins by biting down on the test until it cracks open [23,39], so greater resistance to crushing should make sea urchins less vulnerable to predators [39]. Most studies on the induction of morphological defences in benthic invertebrates have been on gastropod and bivalve molluscs, and the inference that the induction of thicker and/or stronger shells confers some protection against crushing predators is often assumed rather than tested (e.g. [40,41]), but there are some experimental demonstrations (e.g. [42,43]). While we have not experimentally tested this, crush-resistance may be particularly important for small sea urchins, which are generally more susceptible to predation than large individuals, as they are easier to break open [44]. It is important to note that some predators such as larger lobster penetrate the sea urchin through the peristomial membrane [9], so increased crush-resistance alone would not necessarily reduce vulnerability to these predators.

To provide effective protection the sea urchin test must be able to absorb and resist static loads [10], and its strength is usually related to its thickness and porosity [45]. In our six-month-long mesocosm experiment the induction of crush-resistance in juvenile sea urchins through the provision of additional food was mainly due to the infilling of pores, rather than thickening of the test wall. Test thickness was not related to food addition and in fact was greater in the absence of predators. Furthermore, our field survey only found that sea urchin tests were thicker in one of the reserves examined (Leigh) compared to fished sites. These contradictory results for test thickness suggest that test thickness on its own is not a good proxy for overall strength. Indeed, examination of the porosity of the juvenile sea urchin tests indicated that greater strength was achieved through production of denser rather than thicker tests. The relatively weak induction of crush-resistance in response to predation cues in the mesocosm experiment could not be attributed to either an increase in test thickness or the infilling of pores.

In our mesocosm experiment greater food availability resulted in longer-spined individuals, as indicated by the significant food*test diameter interaction. Longer guard spines have been suggested to help protect against predators [9] and sea urchins that lose their guard spines are more susceptible to predation than sea urchins with a full armament of guard spines [46]. There was also a significant interaction between sea urchin size and predation cues in the mesocosm experiment, with larger urchins developing longer spines in the predation cue treatment. While these results suggest that sea urchins may make some investment into developing longer spines when faced with increased food availability and heightened predation risk, there was no difference in spine length between sea urchins collected from reserve and fished reefs. Longer spines are also known to be beneficial for food acquisition [11], and may therefore be expected as an adaptation to food limitation. This was not apparent in our study, suggesting that generating longer spines is energetically uneconomic when food is limiting.

Inducible prey defences are usually considered to be a direct response to the presence of predators, but there is increasing awareness that defences may be induced through indirect mechanisms [2]. While previous studies have shown that changes in foraging activity in response to predators can indirectly induce morphological defences in prey (e.g. [36,37]), to our knowledge the present study provides the first example of predators indirectly inducing morphological defences in prey by increasing the amount of food available to individual prey via a trophic cascade. Access to food resources is fundamental for developing morphological defences, whether these are actively induced [47], or arise incidentally in response to food supply. In the latter case, greater mechanical strength in well-resourced individuals may be a by-product of starved individuals having less resources to allocate to ‘normal’ growth rather than a direct response of well-fed individuals to the threat of predation. For example, crush-resistance of shells in the freshwater snail Mexipyrgus churinceanus is more strongly correlated with the local abundance of a food plant containing materials necessary for shell construction than it is with predation pressure from fish [48]. Therefore, when predators indirectly increase the per capita availability of food to prey, such as through trophic cascades or predation on competitors, prey vulnerability to predation may be reduced due to increased allocation of resources to defences, as we have shown for sea urchins. Similarly, greater food availability may lead to reduced predation risk by increasing body condition and growth rates (e.g. [49]). Indirect induction of morphological defences in prey, mediated by the effect of predators on food availability, may therefore represent an important and previously overlooked mechanism in food webs.

While predators may reduce the abundance or alter the behaviour of prey through trophic cascades [23,50–52], the increased food availability for surviving prey may have a range of benefits including reducing vulnerability to predation. Therefore, indirect induction of morphological defences likely represents an important mechanism in promoting coexistence of predator–prey systems [53] and food chain stability [54].

Ethics

Collection of sea urchins from marine reserves was permitted by the New Zealand Department of Conservation, under research grant 4560.


MATERIALS AND METHODS

Study area for Komodo dragons and ungulates

Extant populations of Komodo dragons, Rusa deer, and wild pigs occur on 4 islands in Komodo National Park (8°35′22″S, 119°36′52″E) in Eastern Indonesia. Populations persist on the 2 large islands of Komodo and Rinca (393.4 and 278.0 km 2 , respectively) and the 2 small islands of Nusa Kode (9.6 km 2 ) and Gili Motang (10.3 km 2 ). A highly seasonal wet–dry cycle is dominant in the region, and there is considerable annual rainfall variation during the short summer monsoon season, which is followed by a long and hot dry season. This climate promotes relatively dry vegetation communities, including open deciduous forest and savannah grassland. Hunting of Komodo dragons and ungulates is prohibited in Komodo National Park, and park rangers regularly patrol all 4 islands. Field data presented in this study were collected within routine long-term Komodo dragon and ungulate population monitoring activities conducted between 2002 and 2015. The focus of our overarching study is to conduct long-term quantitative population assessment of Komodo dragons, ungulates, and habitat quality ( Purwandana et al. 2015 Ariefiandy et al. 2016). The data reported for this specific study, typically represent ancillary research projects conducted at various time points and different spatial scales within the ongoing study. As a consequence, there is temporal separation in completion of the study’s 3 aims. However, our long-term monitoring results suggest that large-scale differences in the population dynamics of Komodo dragons, ungulates, and their habitats have remained relatively stable across the total duration of this study ( Purwandana et al. 2015 Ariefiandy et al. 2016). Thus, experiments conducted at different times within the long-term study are comparable.

Predator–prey habitat use and niche overlap

Over a 6-month period in 2003, we conducted foot-based visual encounter surveys to describe habitat use and niche overlap among Komodo dragons, Rusa deer, and wild pigs. Surveys were conducted at 23 locations comprising 93 km 2 of terrestrial habitats in Komodo National Park. Teams of 6 individuals systematically walked (2–3 km/h) through different habitat types along transects of varying distance pending the topography of the local area. To maximize observations and limit detection bias for each species due to vegetation effects (e.g., structural density of vegetation) obscuring animal sightings, we conducted surveys during the dry season (April–September). During this prolonged period of limited seasonal rainfall, the vegetation structure is reduced due to a die-off in the annual shrub species which increases visibility across the understorey layer ( Auffenberg 1981). Similarly, because many tree species are semi-deciduous or fully deciduous, leaf cover is also considerably reduced at higher vegetation strata during the dry season. These reduced vegetation structural attributes greatly increased our survey visibility and hence detection of large animals across different vegetation communities. In addition, to further increase observations and limit detection bias, we attempted to systematically search the entire land area of each locality to maximize our ability to make direct observations of Komodo dragons, Rusa deer, and wild pigs. This approach was considered more effective than use of a subsampling survey approach at each locality that would instead use a smaller number of randomly positioned transects and culminate in much less survey effort that would be expected to promote reduced animal detections and be more sensitive to survey bias ( Silveira et al. 2003 Ariefiandy et al. 2014). To reduce double counts leading to inflation of observations, observers were restricted to recording animal sightings to within 25 m either side of the immediate position. In total, our visual encounter surveys were conducted over 136 h per individual observer (816 total person observation hours) and covered a distance of 335 km per individual observer (2010 km total person survey distance) to record observations of animals.

At each animal observation for each species we recorded 3 measures of habitat use that could provide a multivariate basis to understand whether Komodo dragon predation risk might affect ungulate prey habitat use. These 3 measures were defined as:

Diurnal activity: We recorded the time of day that the animal was first observed at. Given all animals were observed active, variation in this measure is useful to describe the extent of similarity in species-specific daily activity patterns. Species-specific differences in daily activity patterns could allow ungulate prey to avoid predation risk from Komodo dragons.

Elevational occupancy: We recorded the elevation in meters above sea level that an animal was first observed. The use of elevation provides an index of spatial use in which ungulate could alter their habitat use across an elevational gradient use to minimize risk of Komodo dragon attack.

Vegetation community use: We recorded the dominant vegetation class than an animal occurred in at the time of sighting. The use of vegetation class by animals provides an index of finer-scale spatial habitat use and potentially if different among species, could allow ungulate prey to exploit habitats that minimize risk of attack. We recognized 8 distinct vegetation community types that occur in Komodo National Park these included 1) fore dune vegetation, 2) deciduous monsoon forest, 3) savannah grassland, 4) savannah woodland, 5) closed monsoon forest, 6) ecotone between 2 vegetation communities, 7) mangrove forest, and 8) salt pan vegetation. These vegetation communities differ in plant diversity, dominant species, and structural attributes (e.g., closed forest to open grassland) and are classified in accordance with the definitions used in previous studies ( Auffenberg 1981). These distinct transitions in vegetation type in Komodo National Park occur in response to gradients in water availability and run-off affected by topography or due to differences in vegetation tolerance to soil salinization, salt spray accumulation, and exposure to saltwater inundation ( Auffenberg 1981).

Combined plots of elliptical projections of niche region (NR with individual ellipses of 1000 iterations), density distributions, and raw data scatter plots for each pairwise combination of data estimated for 3 measures of habitat use in Komodo dragons, Rusa deer, and wild pigs. The 3 metrics of habitat use are as follows: 1) Elevation (units = meters above sea level) representing the elevational occupancy in habitat used by each species 2) Diurnal activity time representing the temporal activity use of habitat by each species, and 3) Vegetation community depicting relative species use of the 8 different vegetation communities [1) fore dune vegetation, 2) deciduous monsoon forest, 3) savannah grassland, 4) savannah woodland, 5) closed monsoon forest, 6) ecotone between 2 vegetation communities, 7) Mangrove forest, and 8) Salt pan vegetation] in Komodo National Park.

Combined plots of elliptical projections of niche region (NR with individual ellipses of 1000 iterations), density distributions, and raw data scatter plots for each pairwise combination of data estimated for 3 measures of habitat use in Komodo dragons, Rusa deer, and wild pigs. The 3 metrics of habitat use are as follows: 1) Elevation (units = meters above sea level) representing the elevational occupancy in habitat used by each species 2) Diurnal activity time representing the temporal activity use of habitat by each species, and 3) Vegetation community depicting relative species use of the 8 different vegetation communities [1) fore dune vegetation, 2) deciduous monsoon forest, 3) savannah grassland, 4) savannah woodland, 5) closed monsoon forest, 6) ecotone between 2 vegetation communities, 7) Mangrove forest, and 8) Salt pan vegetation] in Komodo National Park.

To assess if the 3 species demonstrated significant differences in multivariate habitat use (i.e., combined diurnal activity time, elevational occupancy, and vegetation community use), we used nonparametric permutational multivariate analysis of variance (PerMANOVA) using Bray–Curtis dissimilarity index as the distance measure and 1000 permutations ( Oksanen 2011). PerMANOVA was used to test the null hypothesis that the centroids and dispersion of the multivariate species habitat use were significantly similar among the 3 species ( Oksanen 2011).

Niche region dimensions and pairwise niche overlap of species were obtained using the probabilistic method available in the NicheRover package. ( Swanson et al. 2015). First, we used NicheRover to estimate Komodo dragon, Rusa deer, and wild pig niche region dimensions by calculation of bivariate elliptical projections, 1D density plots, and scatterplots. These plots allowed visualization of habitat use similarity for each pairwise comparison of diurnal activity time, elevational occupancy, and vegetation community use for each species. Second, we used NicheRover to utilize a Bayesian probabilistic estimation method for determining bidirectional estimates of pairwise niche region (NR) overlap in multivariate habitat use space among the 3 species. Percentage niche overlap was calculated using respective 95% niche regions between each pair of species. Niche overlap is defined as the probability that an individual from one species is found within the niche region of a second species ( Swanson et al. 2015). Uncertainty in niche overlap was reported as the posterior distribution of overlap percentage along with the Bayesian 95% credible intervals (CIs) for each pairwise species comparison.

Assessing ungulate herd size responses to Komodo dragon predation risk

To estimate whether Komodo dragon predation risk affected Rusa deer and wild pig group size, we evaluated data collected during an annual Komodo dragon and ungulate population sampling undertaken in Komodo National Park. Specifically, we performed distance sampling surveys to evaluate ungulate herd size and density. Concurrently, we also undertook cage trapping to estimate Komodo dragon population densities. These activities were conducted between March and September 2010 at 10 long-term monitoring sites on 4 islands in Komodo National Park.

Four sites were located on Komodo Island: 1) Loh Liang (“K1”), 2) Loh Lawi (“K2”), 3) Loh Sebita (“K3”), and 4) Loh Wau (“K4”). Another 4 sites were located on Rinca Island: 5) Loh Buaya (“R1”), 6) Loh Baru (“R2”), 7) Loh Tongker (“R3”), and 8) Loh Dasami (“R4”). One site was also located on each of the 2 small islands: 9) Gili Motang (“Motang”) and 10) Nusa Kode (“Kode”) ( Supplementary Figure S1 ). Importantly, in recognition that variation in ungulate group size data could be influenced by additional processes beyond Komodo dragon predation risk, we also collected data for additional predictor variables (i.e., ungulate density, Komodo dragon predation risk, habitat risk, and presence/absence of juvenile ungulates in groups). The methods for estimating ungulate group size and predictor variables are described as follows:

Estimation of ungulate group size and population densities

We used distance survey methods to simultaneously estimate spatial variation in ungulate densities, group size, and habitat use and age/sex composition of group. Distance sampling, is a widely used quantitative method to directly estimate ungulate densities ( Thomas et al. 2009). An advantage of the distance sampling method is that it incorporates a detection function, which models the probability of detecting an animal, given its distance from the transect. This method thus has the capacity to compensate and maximize detection probabilities of sighting animals in response to variation in habitat or temporal factors to allow for much improved estimates of animal abundance relative to those obtained from “naive” counts surveys that often do not explicitly consider issues of animal detection probability into abundance estimates ( Buckland et al. 2001).

This study used the same methods from our previously published distance sampling study which is specific to our study system ( Ariefiandy et al. 2013). In brief, distance surveys were conducted in the early morning (06.30–09.30) and late afternoon (15.00–17.30) when ungulates were most active, to increase the likelihood of sighting groups. We surveyed ungulates along 111 transects of variable length (0.5–6.15 km) totaling 163.65 km of surveyed habitat at the 10 site localities. As transects covered the extent of each study site, they enabled sampling across multiple vegetation types where ungulates occur. The same observers (A.A. and D.P.) conducted all surveys. We analyzed data using the program DISTANCE 6.0 release 2 ( Thomas et al. 2009 http://www.ruwpa.st-and.ac.uk/distance/) to estimate species and site-specific density estimates with 95% CI and the coefficient of variation. We used Akaike’s information criterion corrected for small sample sizes (AICc) to assess the relative support for each model. Histograms, quantile–quantile plots and Cramér–von Mises tests were used to assess if data met the assumption of the distance sampling model.

An index of large-scale variation in Komodo dragon predation risk

At the same localities used for observing ungulate group sizes, we estimated Komodo dragon population densities to ascertain landscape-scale variation in Komodo dragon predation risk. Here our rationale was that if the rate of individual ungulate prey encounters with predators affects predation risk, then variation in Komodo dragon density across sites could be proportional to the risk that ungulate face at each locality. To estimate Komodo dragon population density estimates, we used a total of 230 trapping locations (i.e., a fixed point of trapping) at the 10 study sites. These trapping protocols followed previously published methods reporting long-term monitoring and estimation of Komodo dragon population attributes (i.e., phenotypic and demographic attributes Jessop et al. 2006 Purwandana et al. 2014).

In brief, the number of trapping locations was commensurate with the area and vegetation structure of each study site. Within each study site, baited cage traps were placed at individual trapping locations (Lawi, n = 32 Liang, n = 32 Sebita = 21, Wau = 9, Baru = 22, Buaya, n = 22 Tongker, n = 13 Dasami = 24 Motang, n = 16 Kode, n = 12) to capture Komodo dragons. Differences in trap number per site reflected site-specific variation in area and habitat type. Traps comprised purpose built aluminum cage traps (300 cm L × 50 cm H × 50 cm W) fitted with a wire activated front door. The distance between trap locations was set at approximately 500 m to maintain independence among traps. Goat meat (0.5 kg) was used as bait to lure lizards into traps. In addition, a bag of goat meat was suspended 3–4 m above each trap to act as a scent lure to further attract Komodo dragons into each trapping location. At each location, trapping activities occurred over 3 consecutive days, with each trap checked twice daily (8–11 am and 2–5 pm) for the capture of Komodo dragons resulting in 6 sampling events. The time interval between the morning and afternoon daily check for each trap was

6 h. In total this sampling design provided 1380 trapping opportunities for Komodo dragon to be captured.

To estimate site-specific Komodo dragon population density from a single year of cage trapping presence–absence data, we used the Royle–Nichols abundance induced heterogeneity model (henceforth the Royle–Nichols model) in PRESENCE 8.2 ( Hines 2006). The Royle–Nichols model provides estimates of the parameters λ and c, representing average abundance per site and species detectability respectively ( Royle and Nichols 2003). The parameter λ can be interpreted as an index of abundance. However, this assumes that detection of individuals is independent and that site-specific abundance of individuals follows a Poisson distribution (which is the mixture distribution used in PRESENCE models), λ may also be interpreted as the expected number of individuals per sample unit ( Royle and Nichols 2003). We thus divided λ by the sampling site area to estimate average Komodo dragon density at each site ( Ariefiandy et al. 2014). To ensure site-specific estimates of λ were robust, we compared and ranked 6 models using different parameter combinations of k as being either site variant or site invariant (λsite and λ) and c as a function of as being either site variant, survey variant, or site invariant (csite, csurvey, and c). We then used AICc to assess the relative support for each model ( Burnham and Anderson 2003) and used the estimates of λ from the top-ranked model.

Species-specific predation risk

Ungulate group size could be influenced by species-specific susceptibility to Komodo dragon predation. Because Rusa deer are more represented in the Komodo dragon diet relative to their density difference with wild pigs, it suggests that they face higher predation risk ( Auffenberg 1981). We tested for potential species-specific predation risk effecting group size by comparing Rusa deer and wild pig in the same model.

An index of habitat-related predation risk

To consider that ungulates might regulate group size based on finer scale habitat-related predation risk, we scored each group on its use of low-risk habitat or high-risk habitat at the time of observation. These 2 risk categories are based on diurnal observations that suggest that ungulates that utilize open habitats, comprising savannah grassland or savannah woodland, during the day are exposed to low Komodo dragon predation risk. In contrast, ungulates that use closed habitat types, comprised open deciduous forest types or closed dense forest, are at higher risk of Komodo dragon attack. Importantly, low-risk habitats comprise vegetation with limited canopy cover that produces a much hotter microclimate (ground temperatures > 40 °C during midday). Telemetry studies indicate that microclimate causes Komodo dragon to avoid open habitats throughout much of the day ( Imansyah 2006 Harlow et al. 2010). As such these expansive open vegetation communities contain very low Komodo dragon densities and associated predation risk compared with closed habitats that contain much higher densities of Komodo dragons and associated predation risk.

Presence of juveniles

Ungulates can increase group size to offset predation risk to vulnerable new-born or young juvenile offspring. Thus, for each observation of Rusa deer and wild pig, we recorded if individuals within a group contained one or more young juveniles (scored as presence or absence) to ascertain within our model if the presence of juveniles influenced ungulate group size.

Density of ungulates

Group size in ungulates is known to increase as a response to local population density independent of predation risk. To assess whether variation in Rusa deer and wild pig group size was simply an artifact of their own population densities, we considered the effect of ungulate population densities at each study locality. We measured spatial variation in Rusa deer and pig population density and at each of the study sites using distance sampling as detailed above.

To evaluate the effects of Komodo dragon predation risk and other predictor variable on Rusa deer and wild pig herd sizes, we used a full factorial Bayesian linear mixed-effects model (BLMM) with Markov chain Monte Carlo (MCMC) estimation in MCMCglmm, R version 2.15.1 ( Hadfield 2010). To ensure that the parameter estimates were determined by the data, we used uniform prior distributions specified with V = 1 and nu = 0.002 for the variance. Parameter estimates comprised the posterior mode and 95% CIs and were obtained from 1000 iterations subsampled from 650 000 iterations after a 15 000 sample burn-in and a thinning interval of 100, which was sufficient for the MCMC chain to reach stationarity. This model evaluated all additive and interactive effects (i.e., full factorial design) of Komodo dragon predation risk (i.e., Komodo dragon density at each locality), species-specific predation risk (i.e., Rusa deer vs. wild pig), habitat-related predation risk (low-risk open habitat vs. high-risk closed habitat), the effects of presence/absence of juveniles within groups, and the effect of ungulate densities at each study site locality on their respective herd sizes. Site locality was used as random effect in each model. Parameter estimates were considered statistically significant when the 95% CIs did not include 0, and the pMCMC values calculated in MCMCglmm (number of simulated cases that were >0 or <0 corrected for a finite number of MCMC samples) were less than 0.05.

Assessing ungulate antipredator-feeding behavior trade-offs to Komodo dragon predation risk

We conducted field experiments from 6 to 27 July 2015 to determine whether ungulate prey responded to Komodo dragon predation risk by displaying individual-level antipredator behaviors. We undertook experimental field studies that measured the effects of Komodo dragon odor on the foraging behavior of ungulates at feeding stations. The responses of prey to exposure to predator odor cues (e.g., urine, feces, skin/fur swabs) have been widely used to measure the perception of prey to predation risk ( Kats and Dill 1998 Apfelbach et al. 2005). Our design was informed by other reptile predator studies (including Varanid lizards) that have also used odor cues to elicit antipredator responses in mammal or reptile prey ( Carere et al. 1999 Downes 2002 Lloyd et al. 2009 Anson and Dickman 2013).

We placed 48 feeding stations throughout forest habitat at 2 field sites in Komodo National Park. Feeding stations were placed a minimum of 400 m apart to ensure independence of observations among individual ungulates. Each feeding station consisted of a 40-L bowl in which we provided 1000 g of dried corn mixed with 50 g of sea salt and 0.5-L volume of fresh Tamarindus indica leaves. We placed one 10 cm × 10 cm cloth square impregnated with one of 2 odors (Komodo dragon or domestic goat) or a control. The area of the cloth was considered adequate to retain sufficient odor and to also allow ungulates to directly observe the food placed in each bowl. The odors of adult male Komodo dragons (N = 5) and juvenile goats (N = 5) were obtained by swabbing live animals with a cloth soaked in ethanol. The use of ethanol was used to allow the increased concentration and diversity of aqueous and nonaqueous soluble (e.g., lipid) compounds to be absorbed into the cloth squares. Komodo dragon odors were obtained by swabbing the ventral surface of captured animals in the region covering the lower abdomen through to the base of the tail and hind legs (

450 cm 2 ). Swabbing this large area allowed removal of integumental chemicals, fecal and uric acid residue from around the cloaca, and waxy secretions from femoral pores (used in scent marking and other pheromone-based communication of lizards). Thus, this complex composition of this Komodo dragon odor could potentially contain multiple kairomones that could inform ungulates of predation risk ( Apfelbach et al. 2005). Goat odor was obtained by swabbing the pelage of juvenile domestic goats over an equivalent area as used for Komodo dragons. After swabbing, each cloth was stored separately in an air-tight plastic bag and used the same day.

At each feeding station, we fixed a ScoutGuard SG550V camera (HuntingCamOnline, Gadsden, SC) to a nearby tree and focused it on the foraging tray. The camera was programmed to record 1-min videos with 0-s intervals. We located each camera to achieve a 5-m wide field of view centered on the feeding station. We used the video footage that we recorded to quantify the behavior of ungulates during the first 1-min video recorded at each feeding station as a proportion of the entire duration of the individuals’ visit. We deliberately restricted our analysis to the first video at each station because this approach eliminated any variation in the potential food reward present which could influence responsiveness and the influence of previous visitation to foraging stations by conspecifics ( Steindler et al. 2018). By using the first video only, we also reduced the likelihood of quantifying repeated observations from the same individual ( Steindler et al. 2018).

Rusa deer and wild pig behavior recorded at feeding stations was classified into an ethogram comprising 8 behaviors ( Table 1). Here, the proportion of time, or frequency, or presence/absence for each behavior within the first 1 min of video was quantified. We also noted if the visit to the feeding station was made during the day time (after sunrise and prior to sunset) or a nighttime (after sunset and prior to sunrise). We treated all behaviors as mutually exclusive ( Blumstein and Daniel 2007). We scored all videos using the event recorder software JWatcher Video ( Blumstein and Daniel 2007). One observer performed all of the scoring using the JWatcher program. In videos where multiple Rusa deer or wild pigs were present the first individual to appear was scored, but if individuals fought then the video was not used as the intraspecific competition could affect the behavioral analysis and not reflect an individual’s response to the food reward.

Ethogram used to describe and score ungulate behavioral responses to Komodo dragon, goat, or control odors placed at feeding stations

Behavior Description
1. Approach movement at feeding station Total proportion of time within the first minute of observation that an individual spent conducting movement-related approach behaviors at the feeding station. These included ungulates warily approaching the feeding station in a defensive stance, usually in a sideways gait and keeping their body away from the food reward. Or if individuals retreated away from the feeding station.
2. Head-up duration at feeding station Total proportion of time within the first minute that an individual spent at the feeding station with its head up and scanning the surrounding area prior to sniffing or consuming the food reward.
3. Mean head-up duration bout at feeding station The mean time duration of each head-up display bout within the first minute that an individual used at the feeding station to scan the surrounding area prior to sniffing or consuming the food reward.
4. Head-up frequency at feeding station The number of times within the first minute that an individual lifted its head up at the feeding station to scan the surrounding area prior to sniffing or consuming the food reward.
5. Head down at feeding station Total proportion of time within first minute that an individual spent with its head down at the feeding station prior to sniffing or consuming the food reward.
6. Sniffing food bowl at feeding station Total proportion of time within the first minute that an individual spent sniffing and investigating the food reward at the feeding station.
7. Pause sniffing food bowl at feeding station Total proportion of time within the first minute that an individual after sniffing or eating the food reward paused to conduct other behaviors (e.g., head-up vigilance).
8. Consumption of food reward at feeding station The presence or absence of an individual eating the food reward during the first minute of observation.
Behavior Description
1. Approach movement at feeding station Total proportion of time within the first minute of observation that an individual spent conducting movement-related approach behaviors at the feeding station. These included ungulates warily approaching the feeding station in a defensive stance, usually in a sideways gait and keeping their body away from the food reward. Or if individuals retreated away from the feeding station.
2. Head-up duration at feeding station Total proportion of time within the first minute that an individual spent at the feeding station with its head up and scanning the surrounding area prior to sniffing or consuming the food reward.
3. Mean head-up duration bout at feeding station The mean time duration of each head-up display bout within the first minute that an individual used at the feeding station to scan the surrounding area prior to sniffing or consuming the food reward.
4. Head-up frequency at feeding station The number of times within the first minute that an individual lifted its head up at the feeding station to scan the surrounding area prior to sniffing or consuming the food reward.
5. Head down at feeding station Total proportion of time within first minute that an individual spent with its head down at the feeding station prior to sniffing or consuming the food reward.
6. Sniffing food bowl at feeding station Total proportion of time within the first minute that an individual spent sniffing and investigating the food reward at the feeding station.
7. Pause sniffing food bowl at feeding station Total proportion of time within the first minute that an individual after sniffing or eating the food reward paused to conduct other behaviors (e.g., head-up vigilance).
8. Consumption of food reward at feeding station The presence or absence of an individual eating the food reward during the first minute of observation.

Ethogram used to describe and score ungulate behavioral responses to Komodo dragon, goat, or control odors placed at feeding stations

Behavior Description
1. Approach movement at feeding station Total proportion of time within the first minute of observation that an individual spent conducting movement-related approach behaviors at the feeding station. These included ungulates warily approaching the feeding station in a defensive stance, usually in a sideways gait and keeping their body away from the food reward. Or if individuals retreated away from the feeding station.
2. Head-up duration at feeding station Total proportion of time within the first minute that an individual spent at the feeding station with its head up and scanning the surrounding area prior to sniffing or consuming the food reward.
3. Mean head-up duration bout at feeding station The mean time duration of each head-up display bout within the first minute that an individual used at the feeding station to scan the surrounding area prior to sniffing or consuming the food reward.
4. Head-up frequency at feeding station The number of times within the first minute that an individual lifted its head up at the feeding station to scan the surrounding area prior to sniffing or consuming the food reward.
5. Head down at feeding station Total proportion of time within first minute that an individual spent with its head down at the feeding station prior to sniffing or consuming the food reward.
6. Sniffing food bowl at feeding station Total proportion of time within the first minute that an individual spent sniffing and investigating the food reward at the feeding station.
7. Pause sniffing food bowl at feeding station Total proportion of time within the first minute that an individual after sniffing or eating the food reward paused to conduct other behaviors (e.g., head-up vigilance).
8. Consumption of food reward at feeding station The presence or absence of an individual eating the food reward during the first minute of observation.
Behavior Description
1. Approach movement at feeding station Total proportion of time within the first minute of observation that an individual spent conducting movement-related approach behaviors at the feeding station. These included ungulates warily approaching the feeding station in a defensive stance, usually in a sideways gait and keeping their body away from the food reward. Or if individuals retreated away from the feeding station.
2. Head-up duration at feeding station Total proportion of time within the first minute that an individual spent at the feeding station with its head up and scanning the surrounding area prior to sniffing or consuming the food reward.
3. Mean head-up duration bout at feeding station The mean time duration of each head-up display bout within the first minute that an individual used at the feeding station to scan the surrounding area prior to sniffing or consuming the food reward.
4. Head-up frequency at feeding station The number of times within the first minute that an individual lifted its head up at the feeding station to scan the surrounding area prior to sniffing or consuming the food reward.
5. Head down at feeding station Total proportion of time within first minute that an individual spent with its head down at the feeding station prior to sniffing or consuming the food reward.
6. Sniffing food bowl at feeding station Total proportion of time within the first minute that an individual spent sniffing and investigating the food reward at the feeding station.
7. Pause sniffing food bowl at feeding station Total proportion of time within the first minute that an individual after sniffing or eating the food reward paused to conduct other behaviors (e.g., head-up vigilance).
8. Consumption of food reward at feeding station The presence or absence of an individual eating the food reward during the first minute of observation.

We used BLMMs (as specified above) to analyze the effects of time of day, odor type, and their interaction on each behavioral category. As data measuring different ungulate feeding behaviors were drawn from different distributions, we fitted models with their appropriate Gaussian or binomial distributions.


MOVING FORWARD

In a recent review, Sheriff et al. ( 2020 ) emphasised the need to better understand how ecological and environmental context interact with prey responses to predation risk. Focusing on anti-predator behaviour, we address this knowledge gap in two ways. First, our review sheds new light on NCEs by showing when and how contingency can arise from properties of the prey, the predator, and the setting as these effects unfold across three phases (prey risk perception prey responses to perceived risk impacts of these responses on other species Fig. 1). Second, our synthesis of the ‘hunting mode-habitat domain’ and ‘evasion landscape’ concepts offers a unified framework for predicting the form and magnitude of anti-predator behaviour during phase two. Looking ahead, we highlight two knowledge deficiencies that require attention if we are to develop a coherent framework for predicting how NCEs propagate through ecosystems. First, there is insufficient exploration of context-dependent indirect NCEs during phase three. Second, there is need for research focused on the ways in which direct and indirect NCEs are shaped simultaneously, or even interactively, by multiple drivers of context dependence.

Drawing from a broad literature spanning diverse taxa and ecosystems, our review reveals how contingencies in NCEs can arise as a result of many factors. It is hardly surprising, then, that studies have revealed so much variation with respect to whether, and in what way, NCEs manifest in communities (Moll et al. 2016 Gaynor et al. 2019 Prugh et al. 2019 ). We clarify these factors by grouping them into three broad categories: (1) prey properties influencing detection of and responses to risk (2) predator properties shaping their detectability and lethality and (3) properties of the setting influencing the prey’s scope for predator detection and countermeasures. We also emphasise that there is great potential for interplay among them. For example divergent responses to predators with disparate hunting modes could disappear if declining food supply limits prey capacity for defensive investment. Similarly, because prey often have multiple defences whose efficacies are context-specific (Britton et al. 2007 Wirsing et al. 2010 Creel 2018 ), sympatric prey may respond divergently to a shared predator in one setting but similarly in another, depending on the availability of landscape features facilitating particular responses (i.e. the evasion landscape). Moreover, the latter two give rise to an emergent fourth driver, (4) the timing of predation risk, and prey properties then determine how individuals respond to this temporal dimension of danger (Box 3). By implication, predictions based on one driver of contingency, or a single NCE pathway (Preisser & Bolnick 2008 ), may provide an incomplete picture of the impacts of predation risk on prey populations and communities. Rather, examination of NCEs requires thorough consideration of the functional properties of interacting predator and prey species, as well as the circumstances under which these interactions occur (Heithaus et al. 2009 Creel 2011 Schmitz 2017 ). Fortunately, many of these natural history or environmental details are attainable (Wirsing et al. 2010 ), especially given new approaches (e.g. animal-borne video, camera traps, drones) that facilitate placing behavioural data in context (Moll et al. 2007 Wirsing & Heithaus 2014 ).

Our review also highlights the staged manner in which NCE contingencies can manifest. Namely, prey anti-predator investment may vary intra- and inter-specifically as a function of differences in sensory perception (phase one) and the form of any deployed countermeasures (phase two) contingent outcomes during either of the first two phases then determine if, and how, indirect NCEs emerge during phase three. Across taxa, then, prey with greater sensory ability should experience and transmit larger NCEs. Furthermore, the phase in which context dependence arises shapes how the outcome of non-consumptive predator–prey interactions will respond to perturbation. For example landscape changes that reduce prey sensory ability are likely to diminish NCEs, whereas those raising the frequency of encounters with predators by restricting prey habitat domains may elicit increased anti-predator defence during phase two (Schmitz et al. 2004 ) and elevate the potential for indirect NCEs in phase three. Thus, studies exploring phase-specific mechanisms by which prey, predator and landscape properties shape anti-predator investment are crucial to forecasting NCEs in a changing world.

By synthesising the work and concepts of Heithaus et al. ( 2009 ) and Schmitz et al. ( 2017a ), we present a new framework that integrates prey, predator and landscape traits to anticipate the form and magnitude of anti-predator behaviour. This framework is broadly applicable, as evidenced by its ability to retrospectively explain differences in behavioural countermeasures that have been observed in the field across a range of taxa. Consistent with scenario one (Fig. 4c), for example prey species whose habitat domains are nested within those of tiger sharks manifest chronic vigilance and space use that facilitates their escape strategies (Heithaus et al. 2012 ), save when in depressed energetic states (Heithaus et al. 2007 ). Similarly, white-tailed deer whose domains fall within the larger movements of grey wolves exhibit space use changes within their home ranges facilitating their means of predator evasion (Dellinger et al. 2019 ). In contrast, sympatric mule deer practice chronic predator avoidance by shifting to refugia within their domains that are little used by wolves (scenario three Fig. 4b). For both ungulates, the consumptive effects of wolves appear to be limited (Dellinger et al. 2018 ). In the Greater Yellowstone Ecosystem, USA, elk (Cervus canadensis) and wolves have large, overlapping domains, leading to low encounter rates (Cusack et al. 2020 ). Thus, consistent with scenario four (Fig. 4d), elk in this system appear to predominantly experience the consumptive effects of wolves (Peterson et al. 2014 ) and typically exhibit evasive behaviour only during risky times (e.g. Cusack et al. 2020 ). Larger elk survive many encounters with wolves via resistance (Mech et al. 2015 ), further contributing to their tendency to experience consumptive rather than non-consumptive wolf impacts. In an African system with multiple sympatric predators, prey consistently select for habitats offering a lower probability of lethal predator encounters, suggesting that chronic evasive behaviour (under scenarios one and three) may be common where overlapping predator domains preclude outright avoidance (Thaker et al. 2011 ). Accordingly, it underscores characterisation of habitat domains and evasion landscapes as a critical first step in forecasting the extent to which, and how, prey should respond behaviourally to perceived risk during phase two and transmit indirect NCEs in phase three. Our framework also highlights the need to discriminate among prey individuals relying principally on evasion vs. resistance, given that prey expressing the latter group of behaviours are less likely to respond to the threat of predation unless the cue is acute and, consequently, to experience and transmit NCEs. Finally, it gives rise to new hypotheses. For example, in any scenario where predators cannot be avoided spatially and encounters are high enough to warrant anti-predator investment, we might nevertheless expect vigilance and space use that facilitates evasion to relax in prey species that are instead able to avoid the predator(s) temporally (Kohl et al. 2019).

Our survey revealed two knowledge gaps that represent fruitful directions for future research. First, whereas there is ample evidence for context dependence during phases one and two, few studies have rigorously examined contingency in the propagation of indirect NCEs. There are notable examples, including the role of predator hunting mode in shaping indirect NCEs of spiders on plant and soil properties (Schmitz et al. 2017b ), and the impact of prey refugia on indirect non-consumptive relationships between crabs and barnacles (Trussell et al. 2006 ). These studies offer a template for expanded scrutiny of contingencies in NCEs during phase three, which will improve our understanding of when and how predators initiate indirect effects by altering prey traits.

Second, a growing literature underscores the importance of simultaneously considering multiple drivers of contingency in NCEs. For example anti-predator investment by mud crabs varied with their personality (bold vs. shy) and predator hunting mode (actively hunting blue crabs vs. sit-and-wait toadfish, Opsanus tau) (Belgrad & Griffen 2016 ). Thaker et al. ( 2011 ) showed that small members of an African ungulate guild avoided all predators whereas their larger counterparts avoided sit-and-pursue but not active hunters. More work is needed, however, particularly on the importance of three-way interactions among factors drawn from the aforementioned groups.

There are also studies suggesting that interactive impacts of multiple contingent drivers may act collectively to shape indirect NCEs during phase three. For example Murie & Bourdeau ( 2019 ) speculated that, compared to the strong effects initiated by slow-moving sea stars, the absence of direct and indirect non-consumptive effects of crabs and octopuses on snail grazing and kelp, respectively, might owe to the inability of snails to escape these vagile predators. Thus, more mobile prey species with greater scope for avoidance may have responded equivalently to all three predators, yielding similar rather than predator-specific cascades of NCEs. The possibility that interactions between context-dependent factors might modify cascading NCEs has not been tested empirically, however, and thus remains as an exciting research frontier.


Acknowledgments

We thank K. Cheung for volunteering time and resources to capture and transport live sea bass. We also thank K. Schultz, K. O'Brien, B. Moran, J. LaGraff, C. Baillie, and C. Jamison for helping with the experiment. The paper benefitted greatly from the comments of G. Trussell and R. Hughes. Funding for this project was provided, in part, by the Graduate Women in Science Nell Mondy Fellowship, the Switzer Environmental Fellowship, the NOAA Saltonstall Kennedy Grant Program, and crowd funding sourced through Experiment. This is contribution 396 from the Northeastern University Marine Science Center.


An applied ecology of fear framework

Fear, defined here as an animal’s conscious or unconscious perception of risk, is an adaptation that allows an animal to assess the cost of the risk of injury or death (Brown, Laundré, & Gurung, 1999 ). While some have criticized the use of the term “fear” in the context of non-human animals, given that it connotes emotion (Adolphs, 2013 LeDoux, 2014 ), we use it here given that the term has been widely adopted in the ecological literature (Clinchy, Sheriff, & Zanette, 2013 Gaynor et al., 2019 ). Animals differ in how they associate sensory stimuli with perceived risk and respond to that perceived risk, balancing the cost of predation risk with other costs and benefits, including foraging opportunities (Lima, 1986 McNamara & Houston, 1992 Brown, Laundré, & Gurung, 1999 ). This risk assessment and response is mediated by an individual’s state, environmental and social context, and personality (Blumstein & Bouskila, 1996 Lima, 1998 Sih et al., 2004 Owen et al., 2017 ). Acute fear in the presence of an immediate threat can drive reactive anti-predator behavior such as flight and cause physiological stress, while risk assessment in the absence of a direct threat factors into many proactive anti-predator strategies that often have opportunity costs (Lima, 1998 Clinchy, Sheriff, & Zanette, 2013 ). If costly anti-predator behavior compromises survival and reproduction, the risk effects of predation can potentially scale up to alter population dynamics (Lima, 1998 Preisser, Bolnick, & Benard, 2005 Preisser & Bolnick, 2008 ).

Fear can also feed back to shape patterns of predation itself, as anti-predator strategies alter the vulnerability of prey to predators (Sih, 1984 ). Furthermore, by shaping prey behavior and spatiotemporal distribution, fear can alter patterns of herbivory and competition and initiate behaviorally mediated trophic cascades (Schmitz, Beckerman, & O'Brien, 1997 Bucher et al., 2015 ). Despite critical advancements in the understanding of the contributions of fear to ecological dynamics, there remains untapped potential to apply ecology of fear theory to the many animal conservation and management challenges arising due to rapid global change (Berger-Tal et al., 2015 ).

The ecology of fear is relevant to a wide range of animal conservation scenarios. A central goal of many conservation practitioners is to establish and maintain viable populations of wild animal species at a desired density, which can be strongly influenced by risk effects of predation (Creel & Christianson, 2008 ). Conservation practitioners may also wish to change animal behavior and spatiotemporal distribution to shape species interactions (i.e. herbivory, competition) and achieve multi-species conservation goals, to reduce conflict with people (Van Eeden et al., 2018 ) or to enhance people’s positive experiences with wild animals including opportunities for hunting and outdoor recreation (Cromsigt, Kuijper, & Adam, 2013 Larson et al., 2019 ). Wildlife managers may seek to eradicate invasive species or design reintroduction strategies for extirpated native species, and thus increase or decrease predation and mortality rates for a target population (Carthey & Blumstein, 2017 Blumstein, Letnic, & Moseby, 2019 ). Given the importance of risk perception in shaping prey demography, spatiotemporal distribution and predator–prey interactions, managing the ecology of fear can be a powerful tool in each of these contexts.

To effectively apply the ecology of fear concept to conservation science and practice, researchers and practitioners must consider the drivers and consequences of fear, along with the pathways by which fear intersects with management strategies and conservation outcomes (Fig. 1). Ecology of fear dynamics can be directly manipulated to advance particular conservation goals, especially those pertaining to animal behavior and predation rates (e.g. Cromsigt, Kuijper, & Adam, 2013 Bedoya-Perez et al., 2019 ). Alternatively, management practices that are implemented to advance a certain goal may inadvertently reshape the ecology of fear and therefore create new challenges. These cases require a broader consideration of the potential behavioral and physiological responses of target and non-target species to a given management action, along with adaptive management practices. Here, we explore how incorporating the ecology of fear into the management and conservation of wild animal populations could improve outcomes, and point to critical knowledge gaps where future research on fear could inform new conservation strategies.


Is there any example, where cooperative behaviour of predators induce fear in prey population? - Biology

View Issue TOC
Volume 26, No. 3
Pages 156 - 167

The Roles of Large Top Predators in Coastal Ecosystems: New Insights from Long Term Ecological Research

Adam E. Rosenblatt , Michael R. Heithaus, Martha E. Mather, Philip Matich, James C. Nifong , William J. Ripple , Brian R. Silliman
  • Published Online: October 2, 2015
  • Full Article: PDF
  • Export Article Citation: BibTeX | Reference Manager
  • Share

BibTeX Citation

Reference Manager Citation

Article Abstract

During recent human history, human activities such as overhunting and habitat destruction have severely impacted many large top predator populations around the world. Studies from a variety of ecosystems show that loss or diminishment of top predator populations can have serious consequences for population and community dynamics and ecosystem stability. However, there are relatively few studies of the roles of large top predators in coastal ecosystems, so that we do not yet completely understand what could happen to coastal areas if large top predators are extirpated or significantly reduced in number. This lack of knowledge is surprising given that coastal areas around the globe are highly valued and densely populated by humans, and thus coastal large top predator populations frequently come into conflict with coastal human populations. This paper reviews what is known about the ecological roles of large top predators in coastal systems and presents a synthesis of recent work from three coastal eastern US Long Term Ecological Research (LTER) sites where long-term studies reveal what appear to be common themes relating to the roles of large top predators in coastal systems. We discuss three specific themes: (1) large top predators acting as mobile links between disparate habitats, (2) large top predators potentially affecting nutrient and biogeochemical dynamics through localized behaviors, and (3) individual specialization of large top predator behaviors. We also discuss how research within the LTER network has led to enhanced understanding of the ecological roles of coastal large top predators. Highlighting this work is intended to encourage further investigation of the roles of large top predators across diverse coastal aquatic habitats and to better inform researchers and ecosystem managers about the importance of large top predators for coastal ecosystem health and stability.

Citation

Rosenblatt, A.E., M.R. Heithaus, M.E. Mather, P. Matich, J.C. Nifong, W.J. Ripple, and B.R. Silliman. 2013. The roles of large top predators in coastal ecosystems: New insights from long term ecological research. Oceanography 26(3):156&ndash167, https://doi.org/10.5670/oceanog.2013.59.

References

Araujo, M., D.I. Bolnick, and C.A. Layman. 2011. The ecological causes of individual specialisation. Ecology Letters 14:948&ndash958, https://doi.org/10.1111/j.1461-0248.2011.01662.x.

Araujo, M., D.I. Bolnick, G. Machado, A.A. Giaretta, and S.F. dos Reis. 2007. Using &delta 13 C stable isotopes to quantify individual-level diet variation. Oecologia 152:643&ndash654, https://doi.org/10.1007/s00442-007-0687-1.

Bando, K. 2006. The roles of competition and disturbance in a marine invasion. Biological Invasions 8:755&ndash763, https://doi.org/10.1007/s10530-005-3543-4.

Barbier, E., S.D. Hacker, C. Kennedy, E.W. Koch, A.C. Stier, and B.R. Silliman. 2011. The value of estuarine and coastal ecosystem services. Ecological Monographs 81:169&ndash193, https://doi.org/10.1890/10-1510.1.

Beschta, R., and W.J. Ripple. 2009. Large predators and trophic cascades in terrestrial ecosystems of the western United States. Biological Conservation 142:2,401&ndash2,414, https://doi.org/10.1016/j.biocon.2009.06.015.

Binkley, D., M.M. Moore, W.H. Romme, and P.M. Brown. 2006. Was Aldo Leopold right about the Kaibab deer herd? Ecosystems 9:227&ndash241, https://doi.org/󈑖.1007/s10021-005-0100-z.

Bolnick, D., R. Svanback, J.A. Fordyce, L.H. Yang, J.M. Davis, C.D. Husley, and M.L. Forister. 2003. The ecology of individuals: Incidence and implications of individual specialization. The American Naturalist 161:1&ndash28.

Boucek, R., and J.S. Rehage. 2013. No free lunch: Displaced marsh consumers regulate a prey subsidy to an estuarine consumer. Oikos, https://doi.org/10.1111/j.1600-0706.2013.20994.x.

Bruno, J., J.J. Stachowicz, and M.D. Bertness. 2003. Inclusion of facilitation into ecological theory. Trends in Ecology and Evolution 18:119&ndash125, https://doi.org/󈑖.1016/S0169-5347(02)00045-9.

Burkholder, D.A., M.R. Heithaus, J.W. Fourqurean, A. Wirsing, and L.M. Dill. 2013. Patterns of top-down control in a seagrass ecosystem: Could a roving apex induce a behaviour-mediated trophic cascade? Journal of Animal Ecology, https://doi.org/10.1111/1365-2656.12097.

Campbell, M., and F.J. Mazzotti. 2004. Characterization of natural and artificial alligator holes. Southeastern Naturalist 3:583&ndash594, https://doi.org/10.1656/1528-7092(2004)003​[0583:CONAAA]2.0.CO2.

Childers, D. 2006. A synthesis of long-term research by the Florida Coastal Everglades LTER Program. Hydrobiologia 569:531&ndash544, https://doi.org/10.1007/s10750-006-0154-8.

Childers, D., J.N. Boyer, S.E. Davis, C.J. Madden, D.T. Rudnick, and F.H. Sklar. 2006. Relating precipitation and water management to nutrient concentrations in the oligotrophic &lsquo&lsquoupside-down&rsquo&rsquo estuaries of the Florida Everglades. Limnology and Oceanography 51:602&ndash616, https://doi.org/󈑖.4319/lo.2006.51.1_part_2.0602.

Conrad, J., K.L. Weinersmith, T. Brodin, J.B. Saltz, and A. Sih. 2011. Behavioural syndromes in fishes: A review with implications for ecology and fisheries management. Journal of Fish Biology 78:395&ndash435, https://doi.org/󈑖.1111/j.1095-8649.2010.02874.x.

Coulson, R., and T. Hernandez. 1964. Biochemistry of the Alligator. Louisiana State University Press, Baton Rouge, LA.

Craighead, F. 1968. The role of the alligator in shaping plant communities and maintaining wildlife in the southern Everglades. The Florida Naturalist 41:3&ndash7, 69&ndash74.

Croll, D., J.L. Maron, J.A. Estes, E.M. Danner, and G.V. Byrd. 2005. Introduced predators transform subarctic islands from grassland to tundra. Science 307:1,959&ndash1,961, https://doi.org/󈑖.1126/science.1108485.

Dall, S., A.M. Bell, D.I. Bolnick, and F.L.W. Ratnieks. 2012. An evolutionary ecology of individual differences. Ecology Letters 15:1,189&ndash1,198, https://doi.org/󈑖.1111/j.1461-0248.2012.01846.x.

Darimont, C., P.C. Paquet, and T.E. Reimchen. 2009. Landscape heterogeneity and marine subsidy generate extensive intrapopulation niche diversity in a large terrestrial vertebrate. Journal of Animal Ecology 78:126&ndash133, https://doi.org/10.1111/j.1365-2656.2008.01473.x.

Estes, J., K. Crooks, and R. Holt. 2001. Predators, ecological role of. Pp. 857&ndash878 in Encyclopedia of Biodiversity, vol. 4. S. Levin, ed., Academic Press, San Diego, CA.

Estes, J., and D.O. Duggins. 1995. Sea otters and kelp forests in Alaska: Generality and variation in a community ecological paradigm. Ecological Monographs 65:75&ndash100, https://doi.org/󈑖.2307/2937159.

Estes, J., M.L. Riedman, M.M. Staedler, M.T. Tinker, and B.E. Lyon. 2003. Individual variation in prey selection by sea otters: Patterns, causes and implications. Journal of Animal Ecology 72:144&ndash155, https://doi.org/󈑖.1046/j.1365-2656.2003.00690.x.

Estes, J., J. Terborgh, J.S. Brashares, M.E. Power, J. Berger, W.J. Bond, S.R. Carpenter, T.E. Essington, R.D. Holt, J.B.C. Jackson, and others. 2011. Trophic downgrading of planet earth. Science 333:301&ndash306, https://doi.org/󈑖.1126/science.1205106.

Estes, J., M.T. Tinker, T.M. Williams, and D.F. Doak. 1998. Killer whale predation on sea otters linking oceanic and nearshore ecosystems. Science 282:473&ndash476, https://doi.org/󈑖.1126/science.282.5388.473.

Ferretti, F., B. Worm, G.L. Britten, M.R. Heithaus, and H.K. Lotze. 2010. Patterns and ecosystem consequences of shark declines in the ocean. Ecology Letters 13:1,055&ndash1,071, https://doi.org/10.1111/j.1461-0248.2010.01489.x.

Fittkau, E. 1970. Role of caimans in the nutrient regime of mouth-lakes of Amazon affluents (an hypothesis). Biotropica 2:138&ndash142.

Fittkau, E. 1973. Crocodiles and the nutrient metabolism of Amazonian waters. Amazoniana 4:103&ndash133.

Goldschmidt, T., F. Witte, and J. Wanink. 1993. Cascading effects of the introduced Nile perch on the detritivorous/phytoplanktivorous species in the sublittoral areas of Lake Victoria. Conservation Biology 7:686&ndash700, https://doi.org/10.1046/j.1523-1739.1993.�.x.

Hawlena, D., M.S. Strickland, M.A. Bradford, and O.J. Schmitz. 2012. Fear of predation slows plant-litter decomposition. Science 336:1,434&ndash1,438, https://doi.org/󈑖.1126/science.1220097.

Heithaus, M., G.J. Marshall, B.M. Buhleier, and L.M. Dill. 2001. Employing Crittercam to study habitat use and behavior of large sharks. Marine Ecology Progress Series 209:307&ndash310, https://doi.org/10.3354/meps209307.

Heithaus, M., A.J. Wirsing, and L.M. Dill. 2012. The ecological importance of intact top-predator populations: A synthesis of 15 years of research in a seagrass ecosystem. Marine and Freshwater Research 63:1,039&ndash1,050, https://doi.org/10.1071/MF12024.

Heithaus, M., A.J. Wirsing, J.A. Thomson, and D.A. Burkholder. 2008. A review of lethal and non-lethal effects of predators on adult marine turtles. Journal of Experimental Marine Biology and Ecology 356:43&ndash51, https://doi.org/󈑖.1016/j.jembe.2007.12.013.

Helfield, J., and R.J. Naiman. 2006. Keystone interactions: Salmon and bear in riparian forests of Alaska. Ecosystems 9:167&ndash180, https://doi.org/󈑖.1007/s10021-004-0063-5.

Herbert, D., and J.W. Fourqurean. 2008. Ecosystem structure and function still altered two decades after short-term fertilization of a seagrass meadow. Ecosystems 11:688&ndash700, https://doi.org/10.1007/s10021-008-9151-2.

Holtgrieve, G., and D.E. Schindler. 2011. Marine-derived nutrients, bioturbation, and ecosystem metabolism: Reconsidering the role of salmon in streams. Ecology 92:373&ndash385, https://doi.org/10.1890/09-1694.1.

Imhoff, M., L. Bounoua, T. Ricketts, C. Loucks, R. Harriss, and W.T. Lawrence. 2004. Global patterns in human consumption of net primary production. Nature 429:870&ndash873, https://doi.org/10.1038/nature02619.

Jackson, J., M.X. Kirby, W.H. Berger, K.A. Bjorndal, L.W. Botsford, B.J. Bourque, R.H. Bradbury, R. Cooke, J. Erlandson, J.A. Estes, and others. 2001. Historical overfishing and the recent collapse of coastal ecosystems. Science 293:629&ndash638, https://doi.org/10.1126/science.1059199.

Kennedy, C. 2013. Habitat heterogeneity concentrates predators in the seascape: Linking intermediate-scale estuarine habitat to striped bass distribution. MS Thesis, University of Massachusetts, Amherst.

Lancaster, J., M. Dobson, A.M. Magana, A. Arnold, and J.M. Mathooko. 2008. An unusual trophic subsidy and species dominance in a tropical stream. Ecology 89:2,325&ndash2,334, https://doi.org/10.1890/07-0553.1.

Leopold, A. 1943. Deer irruptions. Wisconsin Academy of Sciences, Arts, and Letters 35:351&ndash366.

Maron, J., J.A. Estes, D.A. Croll, E.M. Danner, S.C. Elmendorf, and S.L. Buckelew. 2006. An introduced predator alters Aleutian island plant communities by thwarting nutrient subsidies. Ecological Monographs 76:3&ndash24, https://doi.org/10.1890/05-0496.

Mather, M., J.T. Finn, K.H. Ferry, L.A. Deegan, and G.A. Nelson. 2009. Use of non-natal estuaries by migratory striped bass (Morone saxatilis) in summer. Fishery Bulletin 107:329&ndash338.

Mather, M.E., J.T. Finn, C.G. Kennedy, L.A. Deegan, and J.M. Smith. 2013. What happens in an estuary doesn&rsquot stay there: Patterns of biotic connectivity resulting from long term ecological research. Oceanography 26(3):168&ndash179, https://doi.org/10.5670/oceanog.2013.60.

Mather, M., J.T. Finn, S.M. Pautzke, D. Fox, T. Savoy, H.M. Brundage, L.A. Deegan, and R.M. Muth. 2010. Diversity in destinations, routes and timing of small adult and sub-adult striped bass Morone saxatilis on their southward autumn migration. Journal of Fish Biology 77:2,326&ndash2,337, https://doi.org/󈑖.1111/j.1095-8649.2010.02811.x.

Matich, P., and M.R. Heithaus. 2012. Effects of an extreme temperature event on the behavior and age structure of an estuarine top predator, Carcharhinus leucas. Marine Ecology Progress Series 447:165&ndash178, https://doi.org/10.3354/meps09497.

Matich, P., and M.R. Heithaus. 2013. Multi-tissue stable isotope analysis and acoustic telemetry reveal seasonal variability in the trophic interactions of juvenile bull sharks in a coastal estuary. Journal of Animal Ecology, https://doi.org/󈑖.1111/1365-2656.12106.

Matich, P., M.R. Heithaus, and C.A. Layman. 2011. Contrasting patterns of individual specialization and trophic coupling in two marine apex predators. Journal of Animal Ecology 80:294&ndash305, https://doi.org/󈑖.1111/j.1365-2656.2010.01753.x.

Meserve, P., D.A. Kelt, W.B. Milstead, and J.R. Gutierrez. 2003. Thirteen years of shifting top-down and bottom-up control. BioScience 53:633&ndash646, https://doi.org/󈑖.1641/0006-3568(2003)053[0633:TYOSTA]2.0.CO2.

Myers, R., J.K. Baum, T.D. Shepherd, S.P. Powers, and C.H. Peterson. 2007. Cascading effects of the loss of apex predatory sharks from a coastal ocean. Science 315:1,846&ndash1,850, https://doi.org/10.1126/science.1138657.

Noe, G., D.L. Childers, and R.D. Jones. 2001. Phosphorus biogeochemistry and the impact of phosphorus enrichment: Why is the Everglades so unique? Ecosystems 4:603&ndash624, https://doi.org/10.1007/s10021-001-0032-1.

Ogden, L. 2012. Do animals have personality? BioScience 62:533&ndash537.


Discussion

Even though predators are common stressors in the wild and known to have long-term effects on prey physiology and demography (Boonstra 2013 Clinchy et al. 2013), to our knowledge the indirect consequences of predator presence on individual telomere biology have not been evaluated before. Our results from a wild population of birds facing real predation risk indicate that predators may cause long-term costs in terms of telomere length to their near-living prey. We observed that both male and female parent pied flycatchers nesting near predators (breeding pygmy owls) had shorter telomeres at the end of the chick-rearing period than those nesting at control sites. Moreover, females nesting at owl sites suffered from impaired telomere maintenance during breeding compared to females nesting at control sites. While these results for the parents are correlative because the parents were not randomly allocated to the different environments, they provide the first evidence for a potential predator effect on telomere dynamics that should be verified with manipulative experiments. However, we found no evidence for the hypothesis that predator presence would accelerate telomere shortening in nestlings. Instead, chicks reared at owl sites had consistently longer telomeres during the growth period from day 5 to 12. This suggests that the parents are able to buffer the growth of the chicks against the potential stress caused by predator presence.

Telomere dynamics in parent flycatchers under predation risk

Stress exposure has previously been associated with increased telomere shortening in several species, from human to laboratory and wild animals (Epel et al. 2004 Kotrschal et al. 2007 Herborn et al. 2014 Meillère et al. 2015). In our study, shorter telomeres of pied flycatcher adults nesting at owl sites and the increased telomere shortening in owl-site females may be caused by an increase in glucocorticoids and resulting oxidative stress that arises from the fear and stress of being predated (Angelier et al. 2017). This is supported by the study of Thomson et al. (2010) showing that the levels of stress proteins in blood decreased linearly with the increasing distance to a predator (sparrowhawk Accipiter nisus) nest in the pied flycatcher. Additionally, glucocorticoids have been shown to inhibit telomerase activity (Choi et al. 2008), which could explain the difference in telomere dynamics observed in females between owl and control sites. Indeed, while owl-site females tended to lose telomere length between incubation and chick rearing, females breeding at control sites tended to increase their telomere length. While telomere elongation has been documented in other bird species before (Spurgin et al. 2017), this remains controversial and should be further explored in the future by measuring telomerase activity. Alternatively, telomere lengthening in control-site females could be linked to renewal of blood cells following the first blood sampling. Owl-site females could have less resource to renew their blood cells, which could explain why they show telomere shortening while the telomeres of control-site females show elongation.

A recent hypothesis also suggests that telomere shortening may increase during times of substantially increased energy demands due to specific metabolic adjustments (Casagrande and Hau 2019). For instance, a study on humans found that individuals with high physical activity had shorter telomeres than individuals with moderate physical activity (Ludlow et al. 2008). Pied flycatcher parents have been shown to visit their nests more often under increased predation risk (Hakkarainen et al. 2002 Thomson et al. 2010). The fact that pied flycatcher females at owl sites are lighter than at control sites could be suggestive that they have higher activity levels and, thus, higher energy demands. Therefore, this might contribute to explain the difference observed in telomere dynamics between control- and owl-site females. Although there were no differences in female body mass change between control and owl sites, it is likely that higher activity levels could be induced by predator avoidance even before breeding,਋ut that effects on telomere length might only become visible later on since most of the telomere shortening occurs during the following cellular replication.

We measured male telomere length only once. Therefore, we cannot say with certainty whether the change in telomere dynamics would be the same in males as in females. Nevertheless, similar to the females, males at owl sites had significantly shorter telomere length at the end of chick rearing than control males, which could be the result of faster telomere attrition in owl-site males. However, we cannot exclude the possibility that males in owl sites had already shorter telomeres at the beginning of the breeding season, and that this difference persisted through the study period. It has been shown that pied flycatchers avoid breeding in sites inhabited by a pygmy owl (Morosinotto et al. 2010). Consequently, it is possible that only poor-quality males would have been forced to settle at owl sites, since individuals of good quality may be better in competing over territories, and poor-quality individuals may have initially shorter telomeres than individuals of good quality (Le Vaillant et al. 2015). A potential original quality difference cannot be ruled out for females either. There was however no difference in initial (i.e. during incubation) telomere length or in the change in body mass between incubation and chick rearing between owl- and control-site females. Furthermore, females at control and owl sites had similar clutch size, brood size and managed to raise similar number of fledglings. This data would suggest an absence of difference in quality (at least in terms of breeding performance) for the females breeding at control vs. owl sites. Therefore, in case of an original quality difference between individuals choosing to nest in owl or control sites (in terms of telomere length), it would be sex specific and only concern males. We further attempted to examine for potential quality differences by examining the size of the birds at owl and control sites. At least in pied flycatchers (Potti 1998) and 18 species of Parulidae warblers (Francis and Cooke 1986), males with longer wings have reported to arrive earlier at the breeding grounds, and in other studies early arrival has been linked to potentially better individual quality (Lundberg and Alatalo 1992 Saino et al. 1997 Siitari and Huhta 2002 Smith and Moore 2005 but see Sirkiä and Laaksonen 2009). We did not, however, find any differences in wing length between birds at owl and control sites. Furthermore, in this study, the very first arriving (=𠂟irst breeding) birds were not included, as there were no matching hatching dates in the owl sites to perform the cross-fostering, thus levelling some possible quality differences between birds settling in owl or control sites.

Telomere dynamics in chicks

Telomere shortening is fastest during the growth stage when cell proliferation is high (Spurgin et al. 2017) and accordingly we found a strong reduction in chick telomere length between days 5 and 12. Contrary to our predictions, we observed consistently longer telomeres in chicks reared at owl sites, while there was no significant effect of the site of origin. This suggests that prenatal and early post-natal effects of predator presence (e.g. through transfer of maternal stress hormone) had little or no importance for telomeres, while later post-natal conditions (i.e. after cross-fostering) were more important. Unexpectedly, predator presence during rearing seems to be positive in terms of chick telomere length.

The lack of a prenatal effect was unexpected because stressed females can transfer stress hormones to their developing young, leading to offspring with increased glucocorticoid levels (Saino et al. 2005 Sheriff et al. 2010) and shorter telomeres (Haussmann and Heidinger 2015). In our study, it is possible that there were no differences in maternal glucocorticoid levels between eggs at owl and control sites, or alternatively that the increase of glucocorticoids in the egg was too minor to cause deleterious effects on telomeres. The unexpected apparent positive effect of predator presence on chick telomere length during rearing may be explained by parental behavioural response to the predator threat. Pied flycatcher parents that experience frequent predator encounters resume feeding nestlings quicker than those being less exposed to predators (Thomson et al. 2011) and both Thomson et al. (2010) and Hakkarainen et al. (2002) have reported increased nest visitation and provisioning rates under increased predation risk in pied flycatchers, contrary to what has been found in some other studies (Tilgar et al. 2011 Zanette et al. 2011). Nevertheless, we did not find differences in growth rate between chicks reared at owl and control sites. The potential extra food received by owl-site chicks could have been used for promoting self-maintenance processes (e.g. antioxidant defenses and telomere length maintenance), or parents may have reduced prey load size (Martindale 1982). Carrying food more often to nest could prevent the chicks from begging, which would reduce nest conspicuousness and, although the entrance hole of our nest boxes is too small for the owl to enter, parents may not perceive their chicks being safe as in old natural cavities owls may access the holes by making them larger (Hakkarainen et al. 2002 Thomson et al. 2010). Begging carries an oxidative cost (Moreno-Rueda et al. 2012), thus any reduced begging activity could also contribute to explain the longer telomeres we observe in chicks raised in owl sites. Additionally, chicks that are exposed to nest predator calls can lower their baseline glucocorticoid levels (Ibá༞z-Álamo et al. 2011). High glucocorticoid levels are associated with increased begging rate (Loiseau et al. 2008). Thus, down-regulating glucocorticoids when perceived nest predation risk is high could be adaptive to reduce begging and nest conspicuousness and could contribute to explain our results for chick telomeres. However, gathering data on provisioning rate, prey load size, begging rate and glucocorticoid levels is needed in the future to test these hypotheses.

In conclusion, our study demonstrates that predator presence may affect the telomere length and dynamics of their prey, which could have long-term consequences for the individual in terms of survival probability and add a new hypothesis of how predators may indirectly influence prey demography. While the effects of predation risk seem deleterious in adult birds, the effects seen for nestling telomere length during early rearing were positive, therefore suggesting that different life history stages can be differently affected by increased predation risk. Our results provide further indication for the link connecting environmental stress to cellular/organismal ageing (Angelier et al. 2017), and highlight the potential importance of indirect predator effects on prey physiology for population dynamics.



Comments:

  1. Treasigh

    it seems to me that is the excellent idea

  2. Reid

    I fully share your opinion. This is a great idea. I support you.

  3. Neilan

    Very helpful question

  4. Baram

    It is logical



Write a message