Module 0: What is a Predator?

Before we dissect talons, tail whips and hydraulic jaws, we need a rigorous definition. This opening module fixes the trophic concept of a predator, introduces the functional classification (apex / meso / sub-meso), places the strategy continuum (ambush – sit-and-wait – pursuit) on a common axis, and develops the mathematical backbone of predator–prey ecology: Holling’s functional responses, Lotka–Volterra, the Rosenzweig–MacArthur extension, the paradox of enrichment, trophic cascades, and the keystone species concept.

1. Defining a Predator

A predator is an organism that derives metabolic energy by consuming the biomass of another living organism (the prey) that it has first killed or incapacitated. This trophic definition separates predation from grazing (which rarely kills the resource), parasitism (which typically keeps the host alive), and scavenging(consumption of prey killed by others).

Formally, if N is the prey density and P the predator density, a predation interaction is any process whose consumption rate can be written as a functional response $F(N,P)$ such that the prey death rate contribution is $-F(N,P)\,P$ and the predator energetic intake is$+\varepsilon\,F(N,P)\,P$, where$\varepsilon \in (0,1)$ is the conversion efficiency.

\[\dot N = g(N) - F(N,P)\,P, \qquad \dot P = \varepsilon\,F(N,P)\,P - m\,P\]

where $g(N)$ is the intrinsic prey-growth law and$m$ the predator mortality rate.

Trophic levels

Lindeman’s (1942) trophic-level formalism orders organisms by the number of energy transfers separating them from primary producers. Predators sit at trophic level $\tau \ge 3$. Because only ~10% of energy passes between levels (Lindeman efficiency), top predators are rare and extremely sensitive to perturbations at any lower level.

Functional classification

Apex predators — organisms that, as adults, have no non-human natural predators. Examples: orca, saltwater crocodile, African lion, tiger shark, golden eagle.
Mesopredators — mid-trophic carnivores kept in check by apex predators; they often irrupt when apex predators are removed (Prugh et al., 2009).
Sub-mesopredators — small carnivores (shrews, stoats, wolf spiders) that drive prey dynamics despite modest body size.

2. The Ambush – Sit-and-wait – Pursuit Continuum

Predator hunting strategies do not sort into discrete bins; they occupy a continuous axis driven by the ratio of search time to handling time and by the relative metabolic cost of locomotion versus waiting (Huey & Pianka 1981; Miller et al. 2014).

Three useful anchor points:

Sit-and-wait (web spiders, antlions, pit vipers): negligible search metabolic cost, extreme handling/digestion cost, prey encounter rate limited by prey mobility.
Ambush (jaguar, mantis, crocodile): low-cost stalking punctuated by explosive attack; uses camouflage and burst power.
Pursuit (wolves, cheetah, orca, sailfish): high locomotor cost but high encounter rate; often pack-based to reduce individual cost.

Energetic gradient

The expected net energetic gain per unit time along this continuum obeys

\[\langle E\rangle = \frac{\varepsilon\,E_\text{prey} - c_h\,T_h}{T_s + T_h} - c_s\]

where $T_s$ and $T_h$ are search and handling times, and $c_s, c_h$ are the metabolic costs per unit time. Sit-and-wait minimises $c_s$; pursuit minimises$T_s$.

Hunting-strategy continuum

3. Holling Functional Responses (1959)

Holling’s (1959) disc experiment—where a blindfolded subject picks sandpaper discs off a table—produced a derivation that still underpins modern foraging theory. The predator can be in one of two states: searching(rate $a\,N$ of encounters) or handling (fixed time$h$ per prey). The time budget$T = T_s + h\,F T$ yields the canonical form:

\[F_\text{II}(N) = \frac{a\,N}{1 + a\,h\,N}\]

With limits $F \to a N$ as $N\to 0$ (prey-limited searching) and $F \to 1/h$ as $N\to\infty$(handling-saturated). The three Holling types are:

\[F_\text{I} = a N, \quad F_\text{II} = \frac{a N}{1 + a h N}, \quad F_\text{III} = \frac{a N^2}{1 + a h N^2}\]

Type III produces a sigmoidal response driven by learning or prey-switching, providing low-density prey refuges.

Ratio-dependence and predator interference

When predators interfere, the Beddington–DeAngelis extension gives$F = aN/(1+ahN+wP)$. When prey and predator both scale similarly, Arditi & Ginzburg (1989) propose$F(N/P) = aN/(P+ahN)$, which qualitatively changes the stability structure.

4. Lotka–Volterra Equations and Phase Portraits

Alfred Lotka (1920) and Vito Volterra (1926) independently wrote down the simplest non-trivial predator–prey system, motivated respectively by autocatalytic chemistry and by Adriatic fishery data. With linear prey growth and a type-I response:

\[\dot N = r N - a N P, \qquad \dot P = \varepsilon a N P - m P.\]

The non-trivial fixed point is$N^* = m/(\varepsilon a)$,$P^* = r/a$. Linearisation gives pure imaginary eigenvalues: the system is neutrally stable and orbits in closed curves in phase space. A conserved quantity exists:

\[H(N,P) = \varepsilon a N - m \ln N + a P - r \ln P = \text{const.}\]

Volterra’s principle follows: a uniform harvesting effort that removes both prey and predator at rate $q$ shifts the equilibrium to$N^{*\prime} = (m+q)/(\varepsilon a)$ (more prey) and$P^{*\prime} = (r-q)/a$ (fewer predators), explaining why the suspension of fishing during WWI increased the fraction of predatory fish in Adriatic catches.

5. Rosenzweig–MacArthur & Paradox of Enrichment

Lotka–Volterra ignores self-limitation and assumes an unsaturated response. Rosenzweig & MacArthur (1963) combined logistic prey with a Holling type-II response:

\[\dot N = r N\left(1 - \tfrac{N}{K}\right) - \frac{a N P}{1+a h N}, \quad \dot P = \frac{\varepsilon a N P}{1+a h N} - m P.\]

Setting the derivatives to zero produces the famous nullclines:

\[P_\text{prey}(N) = \frac{r}{a}\left(1-\tfrac{N}{K}\right)(1+a h N), \quad N^* = \frac{m}{a(\varepsilon - m h)}.\]

The prey nullcline is a parabola-like “hump” with a peak at$N_\text{hump} = (K a h - 1)/(2 a h)$. The predator nullcline is a vertical line at $N^*$. A Hopf bifurcation occurs as$N^*$ crosses $N_\text{hump}$: to the right, the equilibrium is a stable spiral; to the left, it is an unstable spiral surrounded by a stable limit cycle.

Paradox of enrichment

Because increasing $K$ pushes the hump to the right while the predator isocline stays put, enrichment (adding nutrients, irrigating, supplementing prey) eventually forces $N^* < N_\text{hump}$ and amplifies oscillations until one species crashes to extinction (Rosenzweig 1971). This is the paradox of enrichment: more food destabilises predator–prey systems.

Simulation 1: Lotka–Volterra & Rosenzweig–MacArthur

Integrates both classical LV and the RM extension. Panel A shows the LV neutrally stable oscillation; panel B its closed phase-plane orbits; panel C the RM nullclines with trajectories for three values of $K$; and panel D scans$K$ to expose the paradox of enrichment as a Hopf transition.

Python

script.py126 lines

import numpy as np
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
from scipy.integrate import odeint

# Classical Lotka-Volterra and Rosenzweig-MacArthur predator-prey dynamics.
# Lotka (1920), Volterra (1926), Rosenzweig & MacArthur (1963).
# LV:
#   dN/dt = r N - a N P
#   dP/dt = e a N P - m P
# RM (logistic prey + Holling type-II functional response):
#   dN/dt = r N (1 - N/K) - a N P / (1 + a h N)
#   dP/dt = e a N P / (1 + a h N) - m P

r = 1.0       # prey intrinsic growth rate (1/time)
a = 0.1       # attack rate
h = 0.5       # handling time
e = 0.3       # conversion efficiency
m = 0.2       # predator mortality
K_values = [5.0, 12.0, 25.0]  # prey carrying capacities for enrichment test

def lv(X, t, r, a, e, m):
    N, P = X
    return [r*N - a*N*P, e*a*N*P - m*P]

def rm(X, t, r, K, a, h, e, m):
    N, P = X
    F = a*N / (1.0 + a*h*N)
    return [r*N*(1.0 - N/K) - F*P, e*F*P - m*P]

t = np.linspace(0, 400, 4000)

# --- Panel 1: LV time series (neutrally stable orbit) ---
sol_lv = odeint(lv, [3.0, 1.0], t, args=(r, a, e, m))

# --- Panel 2: LV phase plane for several initial conditions ---
ICs = [(2.0, 0.5), (3.0, 1.0), (4.0, 1.5), (5.5, 2.0)]

# --- Panel 3: RM phase plane with nullclines + paradox of enrichment ---
# Prey nullcline (dN/dt=0, non-trivial): P = r (1-N/K) (1 + a h N) / a
# Predator nullcline (dP/dt=0, non-trivial): N* = m / (a (e - m h))
N_star = m / (a * (e - m*h))

fig, axes = plt.subplots(2, 2, figsize=(13, 11))
fig.patch.set_facecolor('#0a0a1a')
for ax in axes.flatten():
    ax.set_facecolor('#0a0a1a')
    ax.tick_params(colors='#94a3b8')
    for s in ax.spines.values():
        s.set_color('#334155')
    ax.grid(True, color='#334155', alpha=0.35)

ax1, ax2, ax3, ax4 = axes.flatten()

# Panel 1: LV time series
ax1.plot(t, sol_lv[:, 0], color='#f87171', lw=2, label='Prey N(t)')
ax1.plot(t, sol_lv[:, 1], color='#fb923c', lw=2, label='Predator P(t)')
ax1.set_xlabel('time', color='#94a3b8'); ax1.set_ylabel('density', color='#94a3b8')
ax1.set_title('Lotka-Volterra: neutrally stable oscillation',
              color='#fca5a5', fontsize=13, fontweight='bold')
ax1.legend(facecolor='#1e293b', edgecolor='#334155', labelcolor='#fca5a5')
ax1.set_xlim(0, 200)

# Panel 2: LV phase plane (closed orbits)
for N0, P0 in ICs:
    sol = odeint(lv, [N0, P0], t, args=(r, a, e, m))
    ax2.plot(sol[:, 0], sol[:, 1], lw=1.5, alpha=0.85,
             label=f'(N0,P0)=({N0},{P0})')
# Coexistence point for LV: N*=m/(e a), P*=r/a
N_eq_lv = m/(e*a); P_eq_lv = r/a
ax2.plot(N_eq_lv, P_eq_lv, 'o', color='#fde68a', markersize=10,
         markeredgecolor='#b91c1c', label='fixed point')
ax2.set_xlabel('prey N', color='#94a3b8')
ax2.set_ylabel('predator P', color='#94a3b8')
ax2.set_title('LV phase plane (closed orbits)',
              color='#fca5a5', fontsize=13, fontweight='bold')
ax2.legend(facecolor='#1e293b', edgecolor='#334155', labelcolor='#fca5a5', fontsize=8)

# Panel 3: RM nullclines + trajectories for increasing K
colors_K = ['#fde68a', '#fb923c', '#ef4444']
N_grid = np.linspace(0.01, max(K_values)*1.1, 400)
for K, col in zip(K_values, colors_K):
    # Prey nullcline
    P_null = r * (1.0 - N_grid/K) * (1.0 + a*h*N_grid) / a
    P_null = np.where(P_null > 0, P_null, np.nan)
    ax3.plot(N_grid, P_null, '-', color=col, lw=1.8, label=f'K={K:.0f} prey nullcline')
    # Trajectory
    sol = odeint(rm, [K*0.6, 0.5], t, args=(r, K, a, h, e, m))
    ax3.plot(sol[:, 0], sol[:, 1], '-', color=col, lw=1.2, alpha=0.55)
# Predator nullcline (vertical line at N_star)
ax3.axvline(N_star, color='#22d3ee', lw=2, ls='--', label=f'predator nullcline N*={N_star:.2f}')
ax3.set_xlabel('prey N', color='#94a3b8')
ax3.set_ylabel('predator P', color='#94a3b8')
ax3.set_title('Rosenzweig-MacArthur: nullclines & paradox of enrichment',
              color='#fca5a5', fontsize=13, fontweight='bold')
ax3.legend(facecolor='#1e293b', edgecolor='#334155', labelcolor='#fca5a5', fontsize=8)
ax3.set_xlim(0, max(K_values)*1.05)
ax3.set_ylim(0, 12)

# Panel 4: amplitude of limit cycle vs K (enrichment)
K_scan = np.linspace(2.0, 30.0, 40)
amp_prey = []
amp_pred = []
for K in K_scan:
    sol = odeint(rm, [K*0.6, 0.5], t, args=(r, K, a, h, e, m))
    tail = sol[-1500:]
    amp_prey.append(tail[:, 0].max() - tail[:, 0].min())
    amp_pred.append(tail[:, 1].max() - tail[:, 1].min())
ax4.plot(K_scan, amp_prey, 'o-', color='#f87171', lw=2, label='prey amplitude')
ax4.plot(K_scan, amp_pred, 's-', color='#fb923c', lw=2, label='predator amplitude')
# Hopf bifurcation: K_H = (1 + a h N*) / (a h) when prey nullcline hump crosses N*
K_hopf = (1.0 + a*h*N_star) / (a*h) * (a*h*N_star+1)/(a*h*N_star)
ax4.axvline(K_hopf, color='#22d3ee', ls='--', lw=1.5, label=f'approx. Hopf K~{K_hopf:.1f}')
ax4.set_xlabel('prey carrying capacity K', color='#94a3b8')
ax4.set_ylabel('long-run oscillation amplitude', color='#94a3b8')
ax4.set_title('Paradox of enrichment: amplitude vs K',
              color='#fca5a5', fontsize=13, fontweight='bold')
ax4.legend(facecolor='#1e293b', edgecolor='#334155', labelcolor='#fca5a5', fontsize=9)

plt.tight_layout()
plt.savefig('output.png', dpi=120, bbox_inches='tight', facecolor='#0a0a1a')
print("Lotka-Volterra / Rosenzweig-MacArthur analysis complete.")
print(f"LV coexistence: N*={N_eq_lv:.3f}, P*={P_eq_lv:.3f}")
print(f"RM predator nullcline at N*={N_star:.3f}")
print(f"Enrichment: as K grows past ~{K_hopf:.1f}, oscillation amplitude explodes (paradox).")

Click Run to execute the Python code

Code will be executed with Python 3 on the server

6. Trophic Cascades and the Green-World Hypothesis

Hairston, Smith & Slobodkin (HSS, 1960) asked a deceptively simple question: why is the world green? Their answer—predators keep herbivores below the level that would denude vegetation—defined the green-world hypothesis and launched the study of top-down control.

A trophic cascade is the transmission of predator effects across two or more trophic links. Robert Paine’s (1969) tidepool experiments at Mukkaw Bay removed the sea-star Pisaster ochraceus and watched species richness collapse from 15 to 8 as mussels monopolised the rock—launching the modern keystone species concept.

Subsequent studies provided quantitative case-closed examples:

Estes kelp forests: removal of sea otters by the 19th-century fur trade allowed urchins to graze kelp to “urchin barrens”; reintroduction restored kelp canopies and the entire near-shore community (Estes & Palmisano 1974).
Yellowstone wolf reintroduction(1995–present): wolves suppressed elk browsing of willow and aspen along riparian corridors; the ensuing “landscape of fear” reshaped stream morphology and beaver populations (Ripple & Beschta 2012).
Serengeti lions: removal of apex predators triggered mesopredator release and declines in ground-nesting birds (Prugh et al. 2009).

Three-level trophic cascade

7. Keystone Species and Community Importance

A keystone species exerts a community effect disproportionate to its biomass. Power et al. (1996) formalised keystone community importance as

\[\mathrm{CI}_i = \frac{1}{N-1} \sum_{j\ne i} \left|\frac{t_{j,-i} - t_{j,+i}}{t_{j,+i}}\right|\]

where $t_{j,+i}$ is the trait (abundance, biomass) of species$j$ when species $i$ is present, and$t_{j,-i}$ when absent. High CI per unit biomass = keystone.

Classic exemplars include Pisaster sea stars (Paine 1969), sea otters (Estes & Palmisano 1974), grey wolves (Ripple & Beschta 2012), and ecosystem engineers such as beavers and elephants.

Green-world amendments

Polis (1999) sharpened HSS: the green world is often defended by plant chemistry and toughness rather than by predators alone, and top-down control is strongest in aquatic systems with short food chains (“trophic trickle” vs “trophic cascade”). Meta-analyses (Shurin et al. 2002) confirm stronger cascades in lakes, oceans and streams than in terrestrial systems.

Simulation 2: Holling Type-II Fit & Predator Isocline Geometry

Fits a Holling type-II functional response to synthetic kill-rate data with$\texttt{scipy.optimize.curve\_fit}$, overlays Types I and III, then builds the RM prey-nullcline “hump” and tests whether the predator isocline $N^*$ sits to the left (unstable) or right (stable) of the hump peak.

Python

script.py103 lines

import numpy as np
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
from scipy.optimize import curve_fit

# Holling (1959) functional response types and predator isocline geometry.
# Type I: F = a N            (linear, filter feeders)
# Type II: F = a N / (1 + a h N)  (saturating, most vertebrates)
# Type III: F = a N^2 / (1 + a h N^2) (sigmoid, learning / prey switching)
#
# We (a) fit type II to synthetic predator kill-rate data,
# (b) plot the RM prey nullcline "hump" and show how predator isocline N*
#     intersecting LEFT of the hump destabilises the equilibrium.

# -------- Part 1: fit type-II to synthetic data --------
a_true, h_true = 0.12, 0.6
N_samples = np.array([0.5, 1.0, 2.0, 3.5, 5.0, 7.0, 10.0, 15.0, 22.0])
rng = np.random.default_rng(7)
F_obs = a_true*N_samples/(1.0 + a_true*h_true*N_samples) * (1.0 + rng.normal(0, 0.07, N_samples.size))

def type1(N, a):         return a*N
def type2(N, a, h):      return a*N/(1.0 + a*h*N)
def type3(N, a, h):      return a*N**2/(1.0 + a*h*N**2)

p2, _ = curve_fit(type2, N_samples, F_obs, p0=[0.1, 0.5])
p3, _ = curve_fit(type3, N_samples, F_obs, p0=[0.05, 0.5])

N_fine = np.linspace(0, 25, 400)
F1 = type1(N_fine, p2[0])
F2 = type2(N_fine, *p2)
F3 = type3(N_fine, *p3)

# -------- Part 2: RM prey nullcline hump + predator isocline --------
r, K = 1.0, 20.0
e, m = 0.3, 0.2
a_fr, h_fr = p2[0], p2[1]
# Prey nullcline (non-trivial): P = r(1 - N/K)(1 + a h N)/a
N_grid = np.linspace(0.01, K, 500)
P_null = r*(1.0 - N_grid/K)*(1.0 + a_fr*h_fr*N_grid)/a_fr
# Predator nullcline: N* = m / (a(e - m h))
if (e - m*h_fr) <= 0:
    N_star_iso = K/2  # degenerate; just place in middle
else:
    N_star_iso = m / (a_fr*(e - m*h_fr))
# Hump location: dP/dN = 0 => N_hump = (K a h - 1) / (2 a h)
N_hump = (K*a_fr*h_fr - 1.0) / (2.0*a_fr*h_fr)

# -------- Plotting --------
fig, axes = plt.subplots(1, 2, figsize=(14, 6))
fig.patch.set_facecolor('#0a0a1a')
for ax in axes:
    ax.set_facecolor('#0a0a1a')
    ax.tick_params(colors='#94a3b8')
    for s in ax.spines.values():
        s.set_color('#334155')
    ax.grid(True, color='#334155', alpha=0.35)

ax1, ax2 = axes
# Left panel: type I / II / III
ax1.plot(N_fine, F1, '--', color='#fde68a', lw=1.5, label='Type I (linear)')
ax1.plot(N_fine, F2, '-', color='#f87171', lw=2.5,
         label=f'Type II fit: a={p2[0]:.3f}, h={p2[1]:.3f}')
ax1.plot(N_fine, F3, '-', color='#fb923c', lw=2, label='Type III (sigmoid)')
ax1.plot(N_samples, F_obs, 'o', color='#fca5a5', markersize=7,
         markeredgecolor='#b91c1c', label='observed kills/predator/day')
ax1.axhline(1.0/p2[1], color='#94a3b8', ls=':', lw=1,
            label=f'asymptote 1/h={1/p2[1]:.2f}')
ax1.set_xlabel('prey density N', color='#94a3b8')
ax1.set_ylabel('per-capita kill rate F(N)', color='#94a3b8')
ax1.set_title('Holling (1959) functional responses',
              color='#fca5a5', fontsize=13, fontweight='bold')
ax1.legend(facecolor='#1e293b', edgecolor='#334155', labelcolor='#fca5a5', fontsize=9)

# Right panel: prey nullcline hump + predator isocline
ax2.plot(N_grid, P_null, '-', color='#f87171', lw=2.5, label='prey nullcline (hump)')
ax2.axvline(N_hump, color='#fb923c', ls=':', lw=1.2, label=f'hump top N_hump~{N_hump:.2f}')
ax2.axvline(N_star_iso, color='#22d3ee', ls='--', lw=2,
            label=f'predator isocline N*={N_star_iso:.2f}')
# Shade destabilising region (left of hump)
ax2.axvspan(0, N_hump, alpha=0.08, color='#ef4444')
ax2.text(N_hump*0.5, P_null.max()*0.85, 'UNSTABLE\n(N* left of hump)\ncycles grow',
         color='#fca5a5', ha='center', fontsize=10)
ax2.text(N_hump + (K-N_hump)*0.4, P_null.max()*0.85, 'STABLE\n(N* right of hump)',
         color='#86efac', ha='center', fontsize=10)
ax2.set_xlabel('prey N', color='#94a3b8')
ax2.set_ylabel('predator P', color='#94a3b8')
ax2.set_title('Predator isocline crossing prey hump',
              color='#fca5a5', fontsize=13, fontweight='bold')
ax2.legend(facecolor='#1e293b', edgecolor='#334155', labelcolor='#fca5a5', fontsize=8)
ax2.set_xlim(0, K)
ax2.set_ylim(0, P_null[np.isfinite(P_null)].max()*1.2)

plt.tight_layout()
plt.savefig('output.png', dpi=120, bbox_inches='tight', facecolor='#0a0a1a')
# trapezoid check: mean kill rate across range
mean_F = np.trapezoid(F2, N_fine) / (N_fine[-1] - N_fine[0])
print("Holling fit complete.")
print(f"  Type-II fit a={p2[0]:.3f}, h={p2[1]:.3f}")
print(f"  Mean kill rate over [0,25] = {mean_F:.3f}")
print(f"  Hump peak N_hump = {N_hump:.3f}")
print(f"  Predator isocline N* = {N_star_iso:.3f}")
print(f"  N* < N_hump -> unstable: {N_star_iso < N_hump}")

Click Run to execute the Python code

Code will be executed with Python 3 on the server

8. The Predator Body Plan: Convergent Traits

Across the tree of life, predators converge on a recognisable suite of morphological traits. While the specific implementations differ between mantid shrimp, mako sharks and cheetahs, the functional demands are the same: detect prey, close the gap, kill, and handle.

Forward-facing eyes and high binocular overlap for stereoscopic range estimation. Owls have$\sim 70^\circ$ of binocular overlap; canids$\sim 60^\circ$. Prey species, in contrast, have laterally placed eyes for a wide field of view.
Sharp weaponry: claws, teeth, beaks, chelicerae, raptorial appendages. These scale with prey size rather than predator size: large prey requires disproportionately larger bite force (Wroe et al. 2005).
Locomotor specialisation: burst muscle fibres (type IIx), elongated limbs, flexible spines, or reduced drag profiles in aquatic predators.
Acute distance senses: vision, echolocation, lateral line, infrared pit organs, electrosensation.
Neural investment in predictive motor planning: predators generally have larger relative brain sizes than sympatric prey (Benson-Amram et al. 2016).

The engineering principle is Taylor et al. (1982)’s allometric cost of transport $C_\text{min} \propto M^{-0.32}$: larger predators travel farther per unit energy, enabling the pursuit strategies explored in Module 2.

9. Allometry of Predation and Encounter Theory

In a well-mixed habitat, predator-prey encounter rate follows a kinetic-theory analogue (Gerritsen & Strickler 1977):

\[Z = \pi R^2 N \sqrt{v_p^2 + v_n^2}\]

where $R$ is the detection radius, $N$ is the prey density, and $v_p, v_n$ are predator and prey speeds.

The ratio $v_p/v_n$ is critical. For sit-and-wait predators ($v_p \approx 0$) the encounter rate reduces to$\pi R^2 N v_n$, and so selection pressure rewards expanding the detection cone (web spiders, pit vipers, predatory plants). For pursuit predators (Module 2) the encounter rate grows approximately linearly with$v_p$, driving the evolution of speed.

Body-mass scaling of predation rate

Brose et al. (2006) compiled $> 10\,000$ predator-prey records and showed that the preferred prey/predator body-mass ratio clusters near$10^{-2}$ for invertebrates and $10^{-1}$ for vertebrates. Optimal foraging theory (MacArthur & Pianka 1966) predicts that a predator should include a prey item in its diet if

\[\frac{\varepsilon_i\,E_i}{h_i} > \frac{\sum_{j\in S} a_j N_j (\varepsilon_j E_j - h_j \bar E)}{1 + \sum_j a_j h_j N_j}\]

where $\bar E$ is the current mean intake rate across the current diet $S$. Large, rare prey must yield more energy-per-handling-time to merit inclusion.

Predator / prey body-mass scaling

10. Why Predators Matter: Services and Conservation

Predators are not only ecologically indispensable but also economically valuable. Losey & Vaughan (2006) estimated that invertebrate predators (ladybirds, lacewings, parasitoid wasps) deliver US$\sim \$4.5$ billion per year in pest-suppression services to US agriculture. Large carnivores regulate diseases (coyotes suppressing rodent reservoirs of Lyme disease), control outbreaks of mesopredators, and generate eco-tourism revenue.

Yet predators are disproportionately threatened. Ripple et al. (2014) documented that 77% of the 31 largest mammalian carnivores are in decline, with range contractions of 50–90%. The functional consequence is a global loss of top-down regulation—an important driver of the Anthropocene biodiversity crisis.

10b. Canonical Case Studies

Three case studies anchor the theoretical machinery of this module:

The Hudson Bay lynx–hare cycle

The Hudson’s Bay Company pelt records (Elton & Nicholson 1942) revealed a striking 10-year oscillation between Canada lynx ($Lynx\ canadensis$) and snowshoe hare ($Lepus\ americanus$) densities. Although originally invoked as textbook Lotka–Volterra support, modern analysis (Krebs et al. 2001) shows the oscillation is driven by a tri-trophic interaction with vegetation quality; the hare experiences maternally transmitted stress from predation risk, depressing reproduction for years after a peak. The Hudson Bay data remains the cleanest long-run natural oscillation.

Isle Royale wolves and moose

A continuous 65-year study (Vucetich et al. 2020) on a closed island system in Lake Superior has tracked a two-species predator–prey dynamic under disease shocks (canine parvovirus in 1980) and climate warming. The time series fits a Rosenzweig–MacArthur model with inbreeding depression on the wolves; genetic rescue in 2018 demonstrated that both demographic stochasticity and Allee effects are essential to explain observed crashes.

Australian fox and cat apex-loss

Since 1788, Australia has lost 30+ native mammal species, the worst extinction rate of any continent in the Holocene. Dickman (1996) and Doherty et al. (2016) attribute this to mesopredator release of introduced red foxes and feral cats in the absence of dingoes (which are culled across southern Australia). Where dingoes are protected (north of the dingo fence), native small mammals persist at 10× higher densities. This is a large-scale, accidental natural experiment in top-down control.

Key References

• Lotka, A. J. (1920). “Analytical note on certain rhythmic relations in organic systems.” PNAS, 6, 410–415.

• Volterra, V. (1926). “Fluctuations in the abundance of a species considered mathematically.” Nature, 118, 558–560.

• Holling, C. S. (1959). “The components of predation as revealed by a study of small-mammal predation of the European pine sawfly.” The Canadian Entomologist, 91, 293–320.

• Hairston, N. G., Smith, F. E. & Slobodkin, L. B. (1960). “Community structure, population control, and competition.” American Naturalist, 94, 421–425.

• Rosenzweig, M. L. & MacArthur, R. H. (1963). “Graphical representation and stability conditions of predator–prey interactions.” American Naturalist, 97, 209–223.

• Rosenzweig, M. L. (1971). “Paradox of enrichment: destabilization of exploitation ecosystems in ecological time.” Science, 171, 385–387.

• Paine, R. T. (1969). “A note on trophic complexity and community stability.” American Naturalist, 103, 91–93.

• Estes, J. A. & Palmisano, J. F. (1974). “Sea otters: their role in structuring nearshore communities.” Science, 185, 1058–1060.

• Huey, R. B. & Pianka, E. R. (1981). “Ecological consequences of foraging mode.” Ecology, 62, 991–999.

• Power, M. E. et al. (1996). “Challenges in the quest for keystones.” BioScience, 46, 609–620.

• Polis, G. A. (1999). “Why are parts of the world green?” Oikos, 86, 3–15.

• Ripple, W. J. & Beschta, R. L. (2012). “Trophic cascades in Yellowstone: the first 15 years after wolf reintroduction.” Biological Conservation, 145, 205–213.

• Prugh, L. R. et al. (2009). “The rise of the mesopredator.” BioScience, 59, 779–791.

• Shurin, J. B. et al. (2002). “A cross-ecosystem comparison of the strength of trophic cascades.” Ecology Letters, 5, 785–791.

• Arditi, R. & Ginzburg, L. R. (1989). “Coupling in predator-prey dynamics: ratio-dependence.” Journal of Theoretical Biology, 139, 311–326.

• MacArthur, R. H. & Pianka, E. R. (1966). “On optimal use of a patchy environment.” American Naturalist, 100, 603–609.

• Brose, U. et al. (2006). “Consumer-resource body-size relationships in natural food webs.” Ecology, 87, 2411–2417.

• Wroe, S. et al. (2005). “Bite club: comparative bite force in big biting mammals and the prediction of predatory behaviour in fossil taxa.” Proceedings of the Royal Society B, 272, 619–625.

• Taylor, C. R., Heglund, N. C. & Maloiy, G. M. O. (1982). “Energetics and mechanics of terrestrial locomotion.” Journal of Experimental Biology, 97, 1–21.

• Ripple, W. J. et al. (2014). “Status and ecological effects of the world’s largest carnivores.” Science, 343, 1241484.

• Losey, J. E. & Vaughan, M. (2006). “The economic value of ecological services provided by insects.” BioScience, 56, 311–323.

• Gerritsen, J. & Strickler, J. R. (1977). “Encounter probabilities and community structure in zooplankton.” Journal of the Fisheries Research Board of Canada, 34, 73–82.

• Benson-Amram, S. et al. (2016). “Brain size predicts problem-solving ability in mammalian carnivores.” PNAS, 113, 2532–2537.

• Wilson, A. M. et al. (2013). “Locomotion dynamics of hunting in wild cheetahs.” Nature, 498, 185–189.

• Combes, S. A. et al. (2012). “Linking biomechanics and ecology through predator-prey interactions: flight performance of dragonflies and their prey.” Journal of Experimental Biology, 215, 903–913.

← Course Overview Module 1: Ambush Predators →