Invasion Fitness
Invasion fitness is the per-capita growth rate of a rare mutant introduced into a population dominated by a resident phenotype. It is the central quantitative construct of adaptive dynamics: if invasion fitness is positive, the mutant can establish and potentially replace the resident; if negative, the mutant dies out before reaching appreciable frequency. The invasion fitness landscape — plotting mutant success as a function of both resident and mutant trait values — determines the long-term trajectory of evolution in trait space. The concept generalizes classical relative fitness from population genetics by making fitness explicitly a function of the resident-mutant competition rather than merely environmental parameters.
Invasion Fitness as a Dynamical Systems Concept
Invasion fitness is not merely a population genetics construct. It is a dynamical systems concept that applies to any system in which a rare variant competes with a dominant resident. The sign of invasion fitness determines whether the system is locally stable against perturbations: positive invasion fitness means the equilibrium is unstable, and the mutant will invade; negative invasion fitness means the equilibrium is stable, and the mutant will be repelled. This is the linear stability analysis of evolutionary dynamics, and it is formally identical to the stability analysis of fixed points in ordinary differential equations.
The invasion fitness landscape is a fitness landscape that depends on two variables: the resident trait and the mutant trait. The diagonal of this landscape (where mutant trait equals resident trait) is always zero, because a mutant identical to the resident has the same growth rate. The slope of the landscape perpendicular to the diagonal determines the direction of evolutionary change. If the slope is positive in the direction of increasing trait values, mutants with larger traits will invade, and the population will evolve toward larger trait values. This geometric picture makes invasion fitness a powerful tool for predicting long-term evolutionary trajectories without simulating the full stochastic dynamics.
The Geometry of Invasion and Phase Transitions
The invasion fitness landscape can exhibit complex geometry that produces non-intuitive evolutionary outcomes. A mutant may have positive invasion fitness against one resident but negative invasion fitness against another. This means that the evolutionary outcome depends on the order in which mutants appear — a form of path dependence that is invisible to classical fitness analysis. The system can exhibit evolutionary game theory dynamics: rock-paper-scissors cycles, where each phenotype can invade another but is invaded by a third. These cycles are not exceptions; they are generic features of frequency-dependent selection.
More strikingly, the invasion fitness landscape can undergo qualitative changes as parameters vary — bifurcations in the evolutionary dynamics. A small change in the environment can transform a stable resident into an unstable one, triggering a rapid evolutionary transition. These transitions are the evolutionary analogue of phase transitions in physical systems, and they are characterized by the same mathematical signatures: critical slowing down, increased variance, and sensitivity to perturbations. The adaptive radiation of a species into multiple ecological niches can be understood as a bifurcation in the invasion fitness landscape, where a single stable resident becomes unstable to multiple mutants simultaneously.
Limitations and Extensions
The invasion fitness framework has significant limitations. It assumes that mutants are rare, which means it cannot describe the dynamics when multiple mutants are common simultaneously. It assumes that the population is large, which means it breaks down in small populations where genetic drift dominates. It assumes that the environment is constant, which means it cannot capture the co-evolutionary dynamics where the environment itself evolves in response to the population. These are not minor technicalities; they are fundamental limitations that restrict the framework's applicability.
The extension to finite populations — the stochastic adaptive dynamics framework — shows that invasion fitness is only the leading-order term in a expansion that includes demographic noise, mutational noise, and environmental fluctuations. In small populations, a mutant with negative invasion fitness can still establish through demographic stochasticity, and a mutant with positive invasion fitness can be lost before it reaches appreciable frequency. The deterministic picture of invasion fitness is a large-population approximation, and its failure in small populations is not a correction but a qualitative change in the dynamics.
The extension to co-evolutionary dynamics — where the environment includes other species that are themselves evolving — requires a multi-species invasion fitness concept. A mutant in one species can invade if its per-capita growth rate is positive when all other species are at their equilibrium densities. But if the other species respond evolutionarily to the mutant, the invasion condition changes. The multi-species invasion fitness landscape is a high-dimensional object that is difficult to visualize and even more difficult to measure. Yet it is the true geometry of evolution in ecological communities, and the single-species framework is a projection that loses essential information.
The deepest critique of invasion fitness is that it is a local concept. It tells us whether a rare mutant can invade a resident population, but it does not tell us whether the resulting population is stable against further invasion, or whether the evolutionary trajectory will converge to a fitness maximum, or whether the system will settle into a stable polymorphism. The global dynamics of evolution — the full trajectory through trait space — requires more than local invasion analysis. It requires an understanding of the global geometry of the fitness landscape, and that geometry is rarely known. Invasion fitness is the microscope of evolutionary dynamics: it reveals the local structure with extraordinary precision, but it cannot see the mountain range.