Predatory Behavior

Abstract

Predatory behavior is a form of appetitive behavior: the oriented search, pursuit, and capture by which one animal kills and eats another. The Medical Subject Headings vocabulary indexes it under both appetitive and consummatory behavior—a hunt is an appetitive search that culminates in the consummatory act of the kill. It is among the clearest cases in which a natural behavior can be read on two levels at once: as an economic decision, through optimal foraging theory, which asks when a predator should search, attack, or move on; and as a neural program, through the brain circuits that assemble a hunt from detection to bite. This article sets out the predatory sequence, its economics, the circuits that drive it, and how predators intercept moving prey. A worked example quantifies the marginal value theorem; three demonstrations develop each theme.

Keywords: predatory behavior, optimal foraging, hunting

A cheetah lying in long grass is, for the moment, doing nothing. A herd of gazelle grazes two hundred metres off. The cat watches, picks one animal, and begins a low, slow stalk, freezing whenever a head comes up. At forty metres it breaks into a sprint; the chosen gazelle jinks; the cheetah tracks the turn, closes, and brings it down with a paw to the hindquarters. The whole episode, from the first fixed gaze to the killing bite, is a hunt: a sequence of distinct phases, each with its own rules, strung together into a single goal-directed act of feeding.

This article treats predatory behavior as a structured program that can be analysed at two levels. At the level of behavior, hunting is an economic problem—prey are patchily distributed and costly to catch, and a predator that spends its effort well eats more—which optimal foraging theory formalises. At the level of mechanism, hunting is a neural program in which dedicated circuits in the amygdala and hypothalamus drive the search, the pursuit, and the kill as separable, individually controllable components. Read together, the two levels show how an evolved behavior can be at once a rational strategy and a wiring diagram.

What Predatory Behavior Is

Predatory behavior is the behavior by which a predator locates, attacks, kills, and consumes another animal. In the Medical Subject Headings vocabulary it is indexed under two parents at once—as a form of appetitive behavior, the searching phase of a motivated act, and as a form of consummatory behavior, the completing act itself. The dual placement is not an accident of classification; it reflects the structure of a hunt, which begins as an appetitive search for prey and ends in the consummatory act of the kill and the meal. Predation is therefore the paradigm case of a behavior that spans the whole arc from motivated search to consummation.

Eberhard Curio, in the first book-length synthesis of the subject, treated predation as a behavior to be dissected rather than assumed, setting out its components, the decisions a predator faces at each, and the counter-adaptations of prey (Curio, 1976). The behavior is studied across the animal kingdom because it recurs in forms that are mechanically very different—a toad snapping at a worm, a mantis seizing a fly, a wolf pack running down an elk—yet share an abstract structure. That shared structure is what lets predation serve as a model system: it is concrete enough to measure and general enough that findings in one species illuminate others. Jorg-Peter Ewert made the point at the level of the brain, showing in the toad that a specific, identifiable neural mechanism recognises prey and releases the snapping response (Ewert, 1987).

The Predatory Sequence

A hunt unfolds as an ordered sequence of phases, each governed by different sensory information and different decisions. The predator must first find prey (search), then detect and identify it (detection), then approach without triggering escape (stalk or pursuit), then seize and overcome it (capture and subjugation), and finally consume it. The phases are separable: a predator may detect prey it does not pursue, pursue prey it does not catch, and catch prey in several different ways. The principal stages are summarised in Table 1.

Table 1. The stages of the predatory sequence, each governed by different information and separately modifiable (Curio, 1976; Ewert, 1987).

The detection stage is where the classical ethology of predation is sharpest. Ewert's work on the toad showed that the decision to snap is released by a configural sign stimulus—a dark bar moving in the direction of its long axis reads as “worm” and releases prey-catching, while the same bar moving broadside reads as “non-prey” and does not—and he traced the computation to identified neurons from the retina through the optic tectum (Ewert, 1987). The sequence as a whole is shown in Figure 1.

Optimal Foraging and the Economics of Hunting

Because prey is patchy and costly to catch, how a predator allocates its effort has direct consequences for how much it eats, and natural selection should favour predators that allocate it well. Optimal foraging theory makes this precise by treating the forager as if it maximised its long-term net rate of energy gain. Robert MacArthur and Eric Pianka opened the field by asking how a predator should decide which prey types to pursue and which patches to exploit in a patchy environment, framing both as problems of maximising return per unit of searching and handling time (MacArthur & Pianka, 1966). From this grew the two canonical models: the prey model, which predicts which prey types a predator should attack on encountering them, and the patch model, which predicts how long it should stay in a depleting patch.

Underlying both is the mechanics of capture itself. C. S. Holling's functional response describes how a predator's rate of intake changes with prey density: as prey become more abundant the predator catches more, but the rate saturates, because every capture costs a fixed handling time during which no searching can occur. Holling's disc equation makes this precise and is the origin of handling time as a formal quantity—the very cost that the foraging models trade against search time (Holling, 1959).

The patch problem has an elegant solution. Eric Charnov's marginal value theorem states that a forager in a patch whose intake is slowing should leave at the moment its instantaneous intake rate—the marginal rate—drops to the average rate it could achieve over the whole habitat, counting travel time between patches (Charnov, 1976). The richer the habitat or the shorter the travel between patches, the sooner it should leave each one; the longer the travel, the longer it should stay. Graham Pyke, Ronald Pulliam, and Charnov drew these predictions together and weighed them against the early experimental tests in a review that defined the research programme (Pyke et al., 1977), and David Stephens and John Krebs later gave the theory its mature, book-length statement (Stephens & Krebs, 1986). The first demonstration implements the patch model directly: it plots a diminishing-returns gain curve and finds the residence time at which the predator should abandon the patch for a chosen travel time.

The Neural Control of Hunting

The decisions of the predatory sequence are implemented by identifiable brain circuits, and a decade of work in mice has shown that hunting can be switched on and apart at the level of these circuits. Eliane Comoli and Newton Canteras first mapped the territory by comparing the brain activation of rats hunting insects with that of rats exposed to a live predator, finding that predatory hunting and anti-predator defence drive opposite patterns of activity in the periaqueductal grey—evidence that attacking and being attacked are distinct neural states, not two ends of one arousal axis (Comoli et al., 2003). The lesson that hunting has its own dedicated machinery was then made causal.

Wenfei Han, Ivan de Araujo, and colleagues showed that the central nucleus of the amygdala exerts integrated control over predatory hunting: activating its neurons drove mice to pursue and bite prey and even inanimate objects, while separate projections controlled the pursuit and the kill, so that the sensory and motor components of the hunt could be dissociated (Han et al., 2017). Yi Li and Minmin Luo found a hypothalamic hub for the same behavior, defining circuits in the lateral hypothalamus that drive predation and, through distinct pathways, the evasion of threats (Li et al., 2018). Sae-Geun Park and Daesoo Kim traced a medial preoptic pathway whose activation induces the hunting-like pursuit and seizing of target objects (Park et al., 2018), and Congping Shang and Peng Cao mapped a subcortical excitatory circuit, from the superior colliculus onward, that couples the sight of prey to the launch of a hunt (Shang et al., 2019). Across these studies a common architecture emerges: a sensory trigger, a hub that commits the animal to hunt, and separable outputs for chasing and for the kill. The second demonstration is a schematic of this circuit in which a detection signal and a circuit-gain setting determine whether a modelled hunt proceeds from pursuit to capture.

Prey Capture Across Species

The hardest computational problem in a hunt is the pursuit: catching prey that is itself moving, and often moving evasively. Intercepting a moving target cannot be solved by steering toward where the prey is now, because by the time the predator arrives the prey has moved; the predator must aim at where the prey will be. Different predators solve this prediction problem in strikingly different ways. Matthew Mischiati and Anthony Leonardo reconstructed the three-dimensional flight of hunting dragonflies and showed that the insect steers using an internal model of both its own body and the prey's motion, rotating to keep the prey's image fixed on the eye as it climbs to intercept from below—a predictive strategy, not a simple reaction (Mischiati et al., 2014). Suzanne Kane found that falcons diving on prey follow the trajectory predicted by a guidance law that holds the prey at a constant visual bearing, the same proportional-navigation rule used by guided missiles (Kane & Zamani, 2014).

Vision supplies the information for these pursuits, and its mechanisms are now being dissected. Vivek Nityananda and Jenny Read demonstrated that the praying mantis has genuine stereoscopic depth vision—the only insect known to—using it to judge the distance of moving prey before a strike, a capability they revealed by fitting mantises with tiny 3D glasses (Nityananda et al., 2018). Jennifer Hoy and Cristopher Niell established the laboratory mouse as a model of visually guided prey capture, showing that vision drives the accurate orienting and approach by which a mouse runs down a cricket, and that removing visual input degrades the hunt (Hoy et al., 2016). Not every predator relies on prediction from a distance: Kenneth Catania showed that the tentacled snake exploits its prey's own reflexes, feinting to trigger a fish's escape response and striking not at the fish's current position but at where that reflex will carry it—turning the prey's fastest defence into a liability (Catania, 2009). The third demonstration makes the interception problem concrete: with the prey's speed and the predator's speed set by sliders, the display contrasts a predator that steers toward the prey's present position with one that leads the target.

Worked Example: The Marginal Value Theorem

The patch-leaving rule can be made quantitative. Suppose a predator in a patch accumulates prey at a decelerating rate, because the patch depletes as it hunts. Let the cumulative gain after a time t in the patch be

g(t) = G_max × t / (t + h),

a curve that rises steeply at first and then flattens toward a ceiling G_max, with h setting how quickly it saturates. Between patches the predator must travel for a time τ, during which it gains nothing. Over one patch visit plus the travel to reach it, the long-run rate of gain is

R(t) = g(t) / (τ + t).

The marginal value theorem says the predator should stay until the instantaneous (marginal) intake rate g′(t) falls to this long-run average rate—that is, leave when g′(t) = R(t). For this gain function g′(t) = G_max × h / (t + h)², and setting g′(t) equal to R(t) and simplifying gives a clean result:

t\ = √(h* × τ).

Take a moderately rich patch, h = 2 minutes, and a habitat in which patches are 8 minutes of travel apart, τ = 8. Then t\ = √(2 × 8) = √16 = 4 minutes: the predator should hunt each patch for four minutes before moving on. With a ceiling of G_max = 30 prey items, the gain by then is g(4) = 30 × 4 / (4 + 2) = 20 items, and the realised rate is R(4) = 20 / (8 + 4) = 1.67 items per minute. The marginal rate at that moment, g′(4) = 30 × 2 / (4 + 2)² = 60 / 36 = 1.67 items per minute, is identical, which is exactly the theorem's leaving condition. Now move the patches further apart, τ = 18: the predator should stay longer, t\ = √(2 × 18) = 6 minutes, because the cost of travelling to the next patch has risen. The PatchForagingDemo above is built on this relationship: its travel-time and patch-richness sliders set τ and h, and it reports t\ = √(h* × τ) together with the rate achieved (Charnov, 1976; Stephens & Krebs, 1986).

Discussion

Predatory behavior earns its place in the behavioral sciences because it can be read coherently at two levels that rarely meet so cleanly. At the economic level, optimal foraging theory treats the predator as a rate-maximiser and derives, from that single assumption, quantitative predictions about which prey to attack and how long to stay in a patch—predictions precise enough to be tested and, often, confirmed (Pyke et al., 1977; Stephens & Krebs, 1986). At the mechanistic level, the same behavior resolves into a wiring diagram, a set of circuits in the amygdala and hypothalamus that can be driven to produce pursuit and the kill on demand (Han et al., 2017; Li et al., 2018). The two accounts are complementary: the economic analysis says what the behavior is for and what shape it should take, and the neural analysis says how a brain actually produces that shape.

A second lesson is that hunting is modular. The ethological decomposition of the predatory sequence into separable stages—search, detection, pursuit, capture, consumption—turns out to correspond to something real in the brain, where pursuit and the kill are driven by dissociable projections (Han et al., 2017; Comoli et al., 2003). The interception studies carry a third lesson, about prediction: catching moving prey forces a predator to model the future, and animals as different as dragonflies, falcons, and snakes have each evolved a way to aim not at the prey but at its trajectory (Mischiati et al., 2014; Kane & Zamani, 2014; Catania, 2009). Predation thus connects the economics of decision-making to the neuroscience of action and the computational problem of prediction in a single, observable behavior.

Current Directions

The most active work is the circuit dissection of hunting in the mouse, which has turned predation into a tractable model for how a brain assembles a goal-directed action from parts. Having shown that the central amygdala, lateral hypothalamus, medial preoptic area, and a collicular pathway each contribute, the field is now mapping how these hubs are connected and how a sensory signal is routed through them to commit the animal to a hunt and then to coordinate the chase and the kill as separate outputs (Han et al., 2017; Li et al., 2018; Park et al., 2018; Shang et al., 2019). A recurring theme is the sharp separation between the circuits for attacking and those for being attacked, first seen in the periaqueductal grey and now traced upstream: predation and defence are opposed states with distinct machinery, not points on one continuum of arousal (Comoli et al., 2003; Li et al., 2018).

A second front is the neuroscience and computation of prey capture itself. The mouse prey-capture assay has made the visual control of pursuit accessible to the full toolkit of systems neuroscience, and the question is now how the visual detection of prey is transformed into the motor commands of an accurate approach (Hoy et al., 2016; Shang et al., 2019). In parallel, the discovery that insects as small as the mantis use stereoscopic vision, and that dragonflies hunt with internal models, is reframing interception as a problem solved with surprisingly little neural hardware, and the search is on for the minimal circuits that can compute a predictive pursuit (Nityananda et al., 2018; Mischiati et al., 2014).

Common Misconceptions

Predation is the same as aggression.

It is not. Predatory hunting and defensive or inter-male aggression are driven by different brain circuits: hunting runs through the central amygdala and lateral hypothalamus and is opposed, at the level of the periaqueductal grey, to the circuits for anti-predator defence (Comoli et al., 2003; Han et al., 2017).

A hunt is a single reflexive act triggered by the sight of prey.

A hunt is a sequence of separable stages—search, detection, pursuit, capture, consumption—each governed by its own information and separately modifiable; a predator may detect prey it does not pursue, or pursue prey it does not catch (Curio, 1976).

A predator should always attack any prey it encounters.

Optimal foraging theory shows the opposite: a rate-maximising predator should sometimes ignore an encountered prey type if pursuing it would lower its overall rate of gain, and should leave a patch well before it is exhausted (MacArthur & Pianka, 1966; Charnov, 1976).

Catching moving prey just means chasing toward it.

Steering toward the prey's current position fails, because the prey moves while the predator closes. Successful predators predict the prey's future position—using an internal model, a constant-bearing guidance law, or the prey's own reflexes (Mischiati et al., 2014; Kane & Zamani, 2014; Catania, 2009).

Glossary

Anti-predator defence.

The behavior of an animal avoiding or repelling a predator; driven by brain circuits distinct from, and opposed to, those for predatory hunting.

Appetitive behavior.

The searching, oriented phase of a motivated act that brings an animal into the situation where a consummatory act can occur; the search and pursuit phases of a hunt.

Consummatory behavior.

The completing act of a motivated sequence, such as the capture and eating of prey; the end toward which the appetitive phase is directed.

Fixed action pattern.

A stereotyped, innately programmed sequence of movements released by a specific sign stimulus and, once begun, run to completion; the ethological unit from which Tinbergen built his account of instinctive acts such as the kill.

Handling time.

The time a predator spends subduing, processing, and eating a prey item once caught; it counts against the rate of energy gain in foraging models.

Interception.

The capture of a moving target by aiming at its predicted future position rather than its present one; the core computational problem of pursuit.

Marginal value theorem.

The rule that a forager should leave a depleting patch when its instantaneous intake rate falls to the long-run average rate attainable in the habitat, including travel time.

Optimal foraging theory.

The framework that predicts foraging decisions by assuming an animal acts to maximise its long-term net rate of energy gain.

Patch.

A local concentration of prey or food within a habitat; foraging theory treats a predator's movement as a series of visits to depleting patches separated by travel.

Periaqueductal grey.

A midbrain region surrounding the cerebral aqueduct in which predatory hunting and anti-predator defence drive opposite patterns of activity, marking the two as distinct behavioral states rather than ends of one arousal axis.

Predatory sequence.

The ordered chain of stages—search, detection, pursuit or stalk, capture and subjugation, and consumption—that composes a hunt.

Proportional navigation.

A pursuit strategy that holds the target at a constant visual bearing while closing; used by falcons stooping on prey and by guided missiles alike.

Sign stimulus.

A specific feature of an object—such as a bar moving along its long axis—that releases a fixed response, here the prey-recognition that triggers an attack.

Stereopsis.

Depth perception computed from the disparity between the two eyes' images; used by the praying mantis, uniquely among insects, to judge strike distance to prey.

Key Researchers

Ivan E. de Araujo

(Icahn School of Medicine at Mount Sinai). Neuroscientist who showed that the central nucleus of the amygdala integrates the sensory and motor components of predatory hunting, coordinating pursuit and the jaw-and-neck actions of the kill in mice. Faculty page · Google Scholar

Newton S. Canteras

(University of São Paulo). Neuroanatomist who mapped the hypothalamic and periaqueductal circuits that separate predatory hunting from anti-predator defence, showing the two opposed behaviors drive distinct patterns of neural activation. ORCID

Eric L. Charnov

(University of New Mexico / University of Utah). Evolutionary ecologist whose marginal value theorem supplied optimal foraging theory with its central quantitative result, the rule for when a predator should abandon a depleting patch, and who co-authored the field's defining review. Wikipedia · Google Scholar

Cristopher M. Niell

(University of Oregon). Systems neuroscientist who established the laboratory mouse as a model of visually guided prey capture, showing that vision drives the accurate orienting and approach by which a mouse runs down a cricket. ORCID

Jenny C. A. Read

(Newcastle University). Vision scientist who demonstrated a distinct form of stereoscopic depth vision in the praying mantis, showing that an insect predator uses binocular disparity to judge strike distance to moving prey. ORCID

Nikolaas Tinbergen

(University of Oxford). Dutch-British ethologist and Nobel laureate whose analysis of instinctive behavior in terms of fixed action patterns released by sign stimuli gave the study of hunting its founding framework—the hunt as a chain of innate, stimulus-released acts. Wikipedia

Frequently Asked Questions

What is predatory behavior?

Predatory behavior is the search, pursuit, capture, and consumption of one animal by another for food. It is an ordered sequence of stages, and the Medical Subject Headings vocabulary indexes it as both an appetitive behavior (the searching phase) and a consummatory behavior (the kill and the meal) (Curio, 1976).

How is predation different from aggression?

They are driven by separate brain circuits. Predatory hunting runs through the central nucleus of the amygdala and the lateral hypothalamus, and at the level of the periaqueductal grey it is opposed to the circuits for anti-predator defence, so attacking prey and defending against a threat are distinct neural states rather than degrees of one (Comoli et al., 2003; Han et al., 2017).

What are the stages of a hunt?

A hunt proceeds through search (finding where prey is), detection (recognising and identifying prey), pursuit or stalk (closing the distance), capture and subjugation (seizing and overcoming the prey), and consumption. The stages are separable, and a predator may complete some without the others (Curio, 1976; Ewert, 1987).

What is optimal foraging theory?

It is the framework that predicts a predator's foraging decisions by assuming the animal acts to maximise its long-term net rate of energy gain. From that single assumption it derives which prey types a predator should attack and how long it should exploit a patch before moving on (MacArthur & Pianka, 1966; Pyke et al., 1977).

What is the marginal value theorem?

It is the rule for when to leave a depleting patch: a predator should stay until its instantaneous rate of intake drops to the average rate it could achieve across the whole habitat, counting the travel time between patches. Richer habitats and shorter travel mean leaving each patch sooner (Charnov, 1976; Stephens & Krebs, 1986).

Which brain circuits control hunting?

In mice, activating the central nucleus of the amygdala drives pursuit and biting; the lateral hypothalamus and the medial preoptic area contain circuits that induce hunting; and a subcortical pathway from the superior colliculus couples the sight of prey to the launch of a hunt, with separate outputs for the chase and the kill (Han et al., 2017; Li et al., 2018; Park et al., 2018; Shang et al., 2019).

How do predators catch moving prey?

By predicting where the prey will be, not aiming where it is. Dragonflies steer with an internal model of the prey's motion, falcons hold their target at a constant visual bearing as they dive, and the tentacled snake strikes at where a fish's own escape reflex will carry it (Mischiati et al., 2014; Kane & Zamani, 2014; Catania, 2009).

Why is predation studied as a model behavior?

Because it can be analysed coherently at two levels at once: as an economic decision, through optimal foraging theory, and as a neural program, through the circuits that produce it. The same hunt is both a rate-maximising strategy and a wiring diagram, which makes predation a bridge between the study of decision-making and the neuroscience of action (Hoy et al., 2016; Nityananda et al., 2018).

References