Lotka–Volterra Equations

The Lotka–Volterra equations are two foundational pairs of coupled, first-order, nonlinear ordinary differential equations that govern the continuous-time population dynamics of interacting biological species: the predator–prey model and the interspecific competition model. Formulated independently by the American biophysicist Alfred J. Lotka in 1925 (initially derived from the law of mass action in autocatalytic chemical reactions) and the Italian mathematician Vito Volterra in 1926, these equations established mathematical ecology and dynamical systems analysis of biological food webs.

Volterra's mathematical investigation was inspired by his son-in-law, the marine zoologist Umberto D'Ancona, who analyzed commercial fishery landings from the ports of Trieste, Fiume, and Venice along the Adriatic Sea between 1910 and 1923. D'Ancona observed that the near-cessation of commercial fishing during World War I (1914–1918) caused a pronounced surge in the relative percentage of predatory selachians (sharks, skates, and rays) accompanied by a decline in their prey (food fishes). By translating per-capita birth, mortality, and mass-action encounter rates into differential equations, Lotka and Volterra explained both endogenous predator–prey cycles (such as the 9- to 11-year boreal oscillations of snowshoe hares Lepus americanus and Canadian lynx Lynx canadensis recorded by Hudson's Bay Company fur trappers) and the conditions required for competing species to share an ecological niche.

Dynamical Equations, Equilibria, and Conserved Hamiltonian

Let x(t) denote the population density of prey and y(t) denote the population density of predators at time t. The classical predator–prey model assumes four idealized rules: (1) in the absence of predators, prey follow an exponential growth phase (dx/dt = α*x*, x(t) = x₀ exp(α*t*)) without resource limitation; (2) per-capita prey mortality scales linearly with predator density via mass-action encounters (β*xy*, a Holling Type I functional response); (3) predator birth rate depends directly on the rate of prey consumption converted into new predator offspring (δ*xy*); and (4) in the absence of prey, predators experience exponential decay (y(t) = y₀ exp(−γ*t*)) due to starvation or natural mortality (−γ*y*). The coupled system is:

dx / dt = α*x* − β*xy* = x(α − β*y*)

dy / dt = δ*xy* − γ*y* = y(δ*x* − γ)

where α, β, γ, and δ are positive real rate constants. Setting both derivatives to zero (dx/dt = 0 and dy/dt = 0) yields two equilibria in the (x, y) phase plane: the trivial extinction origin (*x*\*, *y*\*) = (0, 0) (a saddle point), and the interior coexistence equilibrium:

*x*\* = γ / δ , *y*\* = α / β

Notably, the equilibrium prey density *x*\* = γ/δ is determined by the predator's mortality and conversion parameters, whereas the equilibrium predator density *y*\* = α/β is determined by the prey's intrinsic growth and capture parameters. Dividing dy/dt by dx/dt eliminates time t and yields a separable differential equation:

dy / dx = [y(δ*x* − γ)] / [x(α − β*y*)] ⇒ (α / y − β) dy + (γ / x − δ) dx = 0

Integrating both sides produces a conserved first integral (a Hamiltonian-like constant of motion V(x, y) determined by initial conditions x₀, y₀):

V(x, y) = δ*x* − γ ln(x) + β*y* − α ln(y) = constant

Because dV/dt = 0 along every trajectory and V(x, y) is strictly convex with a global minimum at (*x*\*, *y*\*), all non-equilibrium trajectories form closed, counter-clockwise periodic orbits around (*x*\*, *y*\*). Linearizing the Jacobian matrix at (*x*\*, *y*\*) yields purely imaginary eigenvalues λ = ±*i*√(αγ), meaning small-amplitude oscillations cycle with period T ≈ 2π / √(αγ) and the predator peak lags behind the prey peak by one-quarter of a cycle (π/2 radians).

Volterra's Principle and Biocide Resurgence

To explain D'Ancona's Adriatic fishery statistics, Volterra derived Volterra's Principle: over any complete period T of a closed orbit, the time-averaged prey and predator densities ⟨*x*⟩ and ⟨*y*⟩ equal the interior equilibrium values *x*\* = γ/δ and *y*\* = α/β. If an external non-selective harvesting agent—such as commercial fishing nets or broad-spectrum agricultural insecticides—removes both prey and predators at a uniform per-capita rate ε (where ε < α), the governing equations become:

dx / dt = (α − ε)*x* − β*xy* , dy / dt = δ*xy* − (γ + ε)*y*

The shifted time-averaged equilibrium densities become:

⟨*x*⟩_harvest = (γ + ε) / δ > *x*\* , ⟨*y*⟩_harvest = (α − ε) / β < *y*\*

Consequently, indiscriminate harvesting or pesticide application increases the average abundance of the prey (pest) while depressing the predator. Conversely, when fishing paused during World War I (ε → 0), average predatory elasmobranch abundance rose while prey fish abundance declined, confirming D'Ancona's empirical observations and explaining why broad-spectrum agricultural insecticides frequently trigger secondary pest resurgences by decimating predatory coccinellid beetles and parasitoid wasps.

Competition Equations and Coexistence Criteria

When two species N₁ and N₂ compete for shared limiting resources rather than consuming one another, Lotka (1932) and Volterra (1926) extended the logistic growth model incorporating environmental carrying capacity K₁ and K₂ and pairwise competition coefficients α₁₂ and α₂₁ (where α₁₂ measures the per-capita inhibitory effect of species 2 on species 1 relative to the intraspecific effect of species 1 on itself):

dN₁ / dt = r₁N₁ [( K₁ − N₁ − α₁₂N₂ ) / K₁]

dN₂ / dt = r₂N₂ [( K₂ − N₂ − α₂₁N₁ ) / K₂]

Setting the bracketed terms to zero defines two straight zero-net-growth isoclines (ZNGIs) in the (N₁, N₂) phase plane: N₁ + α₁₂N₂ = K₁ and N₂ + α₂₁N₁ = K₂. Depending on the relative magnitudes of the intercepts K₁, K₂, K₁/α₁₂, and K₂/α₂₁, four dynamical outcomes arise, providing the mathematical foundation for Georgii Gause's competitive exclusion principle and modern trophic cascade food-web modules:

Rewriting the Case 4 inequalities in terms of per-capita intraspecific inhibition (α₁₁ = 1/K₁, α₂₂ = 1/K₂) shows that stable two-species coexistence requires niche differentiation such that each species limits its own growth more strongly than it limits its competitor's growth (α₁₂ < K₁/K₂ and α₂₁ < K₂/K₁, which requires α₁₂ · α₂₁ < 1).

Structural Instability and the Rosenzweig–MacArthur Extension

Because the classical predator–prey system possesses a conserved Hamiltonian V(x, y) and purely imaginary eigenvalues at (*x*\*, *y*\*), it is structurally unstable: its neutral cycles depend permanently on initial conditions, and any infinitesimal perturbation flips the system into damped spirals or unbounded oscillations. In 1963, Michael Rosenzweig and Robert MacArthur resolved this defect by introducing two realistic ecological mechanisms: (1) logistic density-dependent self-limitation of the prey with carrying capacity K, and (2) a saturating Holling Type II functional response f(x) = ax / (1 + ahx), which accounts for the finite handling time h a predator requires to pursue, subdue, and digest each prey item:

dx / dt = rx(1 − x / K) − [axy / (1 + ahx)] , dy / dt = [eaxy / (1 + ahx)] − my

Rather than neutrally stable families of orbits, the Rosenzweig–MacArthur model exhibits either a globally stable equilibrium node/spiral at moderate carrying capacity K, or—when nutrient enrichment increases K beyond the stability threshold K > (1 + *ahx*\*) / (ah − a/*x*\*)—undergoes a supercritical Hopf bifurcation into a unique, self-sustaining stable limit cycle whose amplitude is independent of initial conditions. Rosenzweig (1971) termed this destabilization at high K the paradox of enrichment.

Why do predator–prey cycles in the classical Lotka–Volterra model depend on initial population sizes?

The classical Lotka–Volterra predator–prey system is a conservative dynamical system analogous to a frictionless mechanical pendulum, possessing a conserved quantity V(x, y) = δ*x* − γ ln(x) + β*y* − α ln(y). Because there is neither prey self-limitation (carrying capacity K) nor predator saturation (handling time h), trajectories neither spiral inward to a point nor converge to a single limit cycle, instead remaining locked on whichever closed contour of V(x₀, y₀) they began on.

How does Volterra's Principle explain agricultural pest resurgence after spraying insecticides?

According to Volterra's Principle, applying a non-selective mortality rate ε to both a herbivorous insect pest (x) and its natural arthropod predators (y) shifts the average prey density upward to (*x*\* = [γ + ε]/δ) and the average predator density downward to (*y*\* = [α − ε]/β). Because the pest's rapid intrinsic birth rate α outpaces the predator's slower numerical response, killing both guilds releases the pest from top-down control and triggers a secondary pest outbreak.

What is the "paradox of enrichment" in Rosenzweig–MacArthur predator–prey models?

When an ecosystem is fertilized or enriched so that the prey's carrying capacity K rises substantially, the curved prey zero-growth isocline shifts to the right until the vertical predator isocline intersects it to the left of its hump. At this Hopf bifurcation point, the stable coexistence equilibrium becomes unstable and erupts into large-amplitude limit cycles, driving prey and predator densities close to zero at the trough of each oscillation and elevating the risk of stochastic extinction.

1. Lotka, A. J. (1925). Elements of Physical Biology. Williams & Wilkins Company. https://doi.org/10.5962/bhl.title.2633
2. Volterra, V. (1926). "Fluctuations in the Abundance of a Species considered Mathematically." Nature, 118(2972), 558–560. https://doi.org/10.1038/118558a0
3. Rosenzweig, M. L., & MacArthur, R. H. (1963). "Graphical Representation and Stability Conditions of Predator-Prey Interactions." The American Naturalist, 97(895), 209–223. https://doi.org/10.1086/282272
4. Chesson, P. (2000). "Mechanisms of Maintenance of Species Diversity." Annual Review of Ecology and Systematics, 31, 343–366. https://doi.org/10.1146/annurev.ecolsys.31.1.343