Module 2: Pursuit Predators

Grey wolves (Canis lupus) are archetypal endurance pursuit predators. Their hunting mode relies on running prey to exhaustion over distances of 1–5 km at 7–10 m/s (Mech et al. 1998). Unlike cheetahs, wolves are built for sustained aerobic output: type I oxidative muscle fibres, a capacious heart (\(\sim 1.3\%\) of body mass vs. cheetah 0.45%), and blood oxygen-carrying capacity\(\sim 40\%\) above that of comparable canids.

Energetic per-capita cost of locomotion in wolves at pursuit speed is\(C_T \approx 1.4\) J·kg\(^{-1}\)·m\(^{-1}\)(Peterson et al. 2003). For a 40 kg wolf running 3 km this is 168 kJ—a substantial fraction of daily energy needs—so hunts must yield an energetic profit. This sets up the fundamental cost–benefit calculus of pack hunting.

Pack size optimality

MacNulty et al. (2011) showed from 14 years of Yellowstone data that per-capita hunting success in wolves is maximised at intermediate pack sizes (typically 4–5 wolves for elk, 9–13 for bison). Beyond that size, diminishing returns arise because additional wolves free-ride, increasing total prey splitting without adding capture probability. The expected per-capita net energy is:

Landscape of fear

Wolves generate strong non-consumptive effects. Laundré et al. (2001) showed that elk in Yellowstone alter habitat use in response to wolf presence, avoiding riparian zones and thereby releasing willow and aspen (Ripple & Beschta 2012). This “landscape of fear” is one of the clearest demonstrations that pursuit predation affects prey populations not only through kills but through behavioural cascades.

The cheetah holds the land-animal speed record: top speeds of 105 km/h with peak acceleration\(\sim 10\) m/s\(^2\)—exceeding every production sports car (Wilson et al. 2013). Botswana GPS + accelerometer collars recorded chases up to 120 m, with typical strides of\(7\text{--}9\) m enabled by spinal flexion that effectively adds a third stride phase to each cycle.

Traction and the unique claws

Cheetahs are the only cats with non-retractable claws. These function as cleats, giving the traction needed for high-speed turns and accelerations on grass and sand (Patel & Braae 2013). The traction-limited maximum drive force is:

Drag-dominated top speed

At top speed the drag force dominates:

Taylor’s thermal-limit hypothesis

Taylor & Rowntree (1973) and Taylor (1974) famously argued that cheetahs abandon \(\sim 30\%\) of chases because muscle heat generation exceeds cooling capacity. Body temperature rises at\(\sim 0.6\) °C per 100 m sprint; the\(40.5\) °C threshold corresponds to the point where cytochrome c oxidase begins to lose integrity. Recent work (Hetem et al. 2013) challenges the strict thermal-limit interpretation, showing field body temperatures rarely exceed 39.5 °C and suggesting that abandonment is partly behavioural (avoiding risk of injury from hoofed prey). Our Simulation 1 captures both effects: drag-limited speed and heat-limited duration.

Tail as rudder

Patel & Braae (2013) modelled the cheetah tail as an inertial rudder. For a moment of inertia \(I_\text{tail}\approx 0.03\) kg·m\(^2\), rotating the tail with angular velocity \(\dot\theta\)generates an opposing body yaw \(\dot\phi = -I_\text{tail}/I_\text{body}\,\dot\theta\). This allows the cheetah to turn while maintaining forward traction, the mechanism underlying the characteristic “flick-and-cut” turns captured in the Wilson et al. (2013) data.

Orcas are the ocean’s apex predators, coordinating pod attacks on prey ranging from herring (Norwegian resident pods) to blue whales (Ness Ridge “killer pod” observations). Cruising speed is\(\sim 10\) km/h; short-burst pursuit up to 56 km/h (Similä & Ugarte 1993).

Coordinated strategies include carousel feeding (herring ball formation and stunning tail slaps), wave washing (generating a bow wave to knock seals off ice floes in Antarctic “pack ice ecotype” B1), and strand feeding (Valdes Peninsula pods beaching temporarily to grab sea-lion pups). These behaviours are culturally transmitted and differ sharply between ecotypes even in sympatry.

Propulsive efficiency in delphinids reaches \(\sim 0.85\), near the theoretical hydrodynamic optimum (Fish 1998). Hydrodynamic drag on a cetacean scales as \(F_d = \tfrac{1}{2}\rho C_D S v^2\) with\(C_D \approx 0.003\) (laminar-flow achievement) and\(S\) the wetted surface area.

The sailfish (Istiophorus platypterus) is frequently cited as the fastest fish at burst speeds of 110 km/h, although direct measurements (Marras et al. 2015) give a more conservative maximum burst of \(\sim 36\) km/h during sardine “slashing” attacks. Related billfishes (marlin, spearfish) employ their elongated rostrum (bill) to slash through schooling prey, stunning several fish per pass.

The rostrum provides a hydrodynamic advantage: Domenici et al. (2014) used high-speed video to show that the bill creates a low-pressure wake behind the head that reduces pressure gradients on the prey’s lateral line, suppressing escape responses. Slashing accuracy improves when the billfish attacks from the trailing edge of a bait-ball, exploiting reduced prey turning capacity.

Acceleration scaling

Small fish accelerate faster than large fish:\(a_\text{max} \propto M^{-0.33}\) (Webb 1983). Sailfish bursts reach peak accelerations of 125 m/s\(^2\) for\(\sim 100\) ms during the slash attack—comparable to the mantid’s raptorial appendage, scaled to a 40 kg fish.

Peregrine falcons (Falco peregrinus) execute the stoop—a diving attack from several hundred metres altitude. Radar-tracked dives have recorded airspeeds of \(\sim 320\) km/h (Tucker 1998), the fastest known predator speeds. The terminal velocity for a streamlined falcon is:

Peregrines curve their trajectory inward during the stoop rather than executing a direct line. Ponitz et al. (2014) showed this spiral pursuit keeps the prey in the line of one eye while minimising head-turn drag; the optimal attack trajectory is a logarithmic spiral, consistent with motion camouflage strategies used by insects (dragonflies) and fighter pilots alike.

At pursuit speeds Reynolds numbers (\(\text{Re} = \rho v L/\mu\)) are typically \(10^5\text{--}10^7\)—fully turbulent flow, so drag force scales as\(F_d \propto v^2\) and required power\(P \propto v^3\). Doubling top speed requires an eight-fold increase in mechanical power output.

The allometric consequence (Bejan & Marden 2006, Garland 1983) is a characteristic speed-mass curve with a maximum at\(M \approx 100\) kg—precisely where cheetah, greyhound, pronghorn and Thomson’s gazelle all cluster. Below this mass, limb-stride frequency limits speed; above it, tendon and muscle strength cannot grow fast enough to overcome quadratic drag.

Pack hunting creates a coordination problem. If cooperation increases capture probability as\(p(N) = 1 - (1 - p_1)^N\) (independent hunter assumption), then per-capita yield is\(E_\text{prey}(1 - (1-p_1)^N)/N\), which has a unique maximum at small \(N\). Creel & Creel (1995) found this is precisely what African wild dog (Lycaon pictus) packs achieve when intra-pack competition is strong.

When coordination genuinely synergises (packs surround prey and cut off escape), capture probability can rise super-additively, breaking the independent-hunter scaling. Fanshawe & FitzGibbon (1993) estimated wild dog cooperative synergy at \(\sim 20\%\) above independent expectation. The resulting optimal pack size depends strongly on prey body mass—which Simulation 2 makes explicit.

Aerobic vs anaerobic budgets

The basic dichotomy of pursuit modes is:

• Aerobic / endurance: wolves, African wild dogs, hyenas, humans. Heat load managed by sweat (humans) or panting. Cost \(\sim 1.4\) J/kg/m; sustained up to hours.
• Anaerobic / sprint: cheetahs, greyhounds. Phosphagen + glycolytic pathways sustain output for\(\sim 30\) s before lactate accumulates. Cost per meter lower but cumulative heat and H\(^+\) load force a ceiling.
• Ambush-pursuit hybrid: lions, tigers, leopards. Stalk to \(< 30\) m then sprint\(\sim 50\text{--}100\) m. Balanced budget.

Pursuit predation is an evasive-target control problem. The classic formulation (Howland 1974) considers a predator at speed\(v_p\) chasing prey at speed \(v_n\) with maximum centripetal accelerations \(a_p, a_n\). If\(v_p > v_n\) but \(a_n > a_p\), the prey can escape by making tight turns that the predator cannot follow.

For uniform circular motion, the achievable turn radius is\(R_\text{min} = v^2/a\). If prey and predator reach\(R_\text{min}^\text{prey} < R_\text{min}^\text{pred}\), the prey escapes. Many prey species (Thomson’s gazelle, bobwhite quail, fruit flies) maintain speeds below their own\(v_\text{max}\) specifically to keep turn radius low.

Protean behaviour

When the predator is highly capable, prey may exhibit protean (unpredictable) behaviour, randomising their trajectory to exceed the predator’s ability to predict. Jones et al. (2011) quantified this in zebrafish and showed that protean trajectories are statistically indistinguishable from a random walk, within the flow-fluid constraints.

• African wild dog (Lycaon pictus): endurance runner on open savanna; \(\sim 80\%\) hunt success, the highest reliably measured of any large predator. Relies on coordinated disruption of prey herds and vocal communication (Walker et al. 2017).
• Spotted hyena (Crocuta crocuta): matriarchal clans pursue zebra and wildebeest over distances of 3–5 km; metabolic scope 10× basal.
• Dolphin (Tursiops) “fish-whacking”: use echolocation to drive fish into shallow water, then tail-slap to stun; culturally transmitted in Shark Bay, Australia.
• Dragonfly (Anax) aerial interception: uses visual prediction to intercept prey rather than pursue them, achieving 95% strike success (Combes et al. 2012). Functionally a “flying ambush” disguised as pursuit.
• Human persistence hunting: ancestral Homo exploited thermoregulatory advantage (sweat) to out-endure ungulates in open savanna (Bramble & Lieberman 2004). This is the only pursuit strategy where the key trait is heat dissipation, not speed.

Pursuit predators disproportionately structure ecosystems through density-mediated and trait-mediated effects. Density mediation is the classic numerical response: wolves eat elk, so elk density falls. Trait mediation is subtler: prey alter habitat use, group size, vigilance and timing in response to perceived risk, changing their ecological footprint even without being killed.

Schmitz (1998) measured relative contributions in grassland arthropod food webs and found that trait-mediated effects often exceeded density-mediated ones. For Yellowstone wolves, Ripple & Beschta (2012) attribute much of the willow/aspen recovery to behavioural avoidance by elk rather than to population reduction.

Giving-up density

Brown’s (1988) giving-up density (GUD) framework quantifies the residual resource level at which prey abandon a patch due to predation risk. GUD is a direct measure of the “cost of fear”:

All terrestrial pursuit predators transition between gaits as speed increases: walk, trot, canter, gallop. At each transition, metabolic cost of transport drops (Hoyt & Taylor 1981), producing the characteristic U-shaped efficiency curve. The predator’s gallop is the only gait that allows simultaneous airborne phases, enabling spinal storage and release of elastic strain energy.

Spinal flexion

The cheetah’s \(\sim 80\)° spinal flexion per stride adds \(\sim 1\) m to the effective stride length. Hudson et al. (2011) used marker-based kinematics to show that the horizontal component of back motion accounts for 40% of propulsion at top speed—an extraordinary fraction compared to human running (\(< 5\%\) back contribution). Greyhounds share this trait; horses, with their rigid backs, do not.

Tendon spring storage

Achilles-type tendons store \(\sim 35\%\) of the kinetic energy at footfall and return it during push-off (Alexander 1984). Fast runners (cheetah, greyhound, kangaroo) have long, thin tendons with near-optimal elastic energy storage coefficients.

In the canonical Huey & Pianka (1981) framing, the two foraging strategies differ along a continuous axis of\(c_s / v_p\): ratio of search-cost to speed. Pursuit predators pay high \(c_s\) and need high \(v_p\); ambushers minimise both but accept long waiting. The optimal strategy is density-dependent: when prey density is high, ambush wins; when prey density is low and mobile, pursuit wins.

Lima & Bednekoff (1999) added the predation-risk allocation hypothesis: prey modulate vigilance based on predator strategy, with pursuit predators generating chronic low-level vigilance while ambushers trigger pulse vigilance at specific microsites. This feeds back into predator strategy choice, producing evolutionary lock-in.