Home > Positioning > Persons > Wright

Sewall Wright (1889–1988)

Wright was a geneticist and evolutionary theorist whose work, alongside Fisher’s and Haldane’s, founded mathematical population genetics. Where Fisher emphasised large panmictic populations and the directional power of natural selection, Wright emphasised population structure — small, partially isolated populations in which genetic drift, epistasis, and migration interact with selection to produce evolutionary change. His adaptive landscape metaphor became one of the most influential conceptual tools in evolutionary biology. His shifting balance theory — proposing that evolution proceeds most effectively when drift moves small populations between adaptive peaks and migration spreads successful combinations — structured population-genetic debate through the twentieth century and remains contested.

Life

Born 21 December 1889 in Melrose, Massachusetts. Grew up in a family of scholars — his father taught economics, his brother Quincy Wright became a prominent political scientist. BS from Lombard College, Illinois (1911). ScD from Harvard (1915), working on coat colour genetics in guinea pigs under William Ernest Castle. Joined the US Department of Agriculture as senior animal husbandman (1915–25), where his work on inbreeding and crossbreeding in guinea pig populations — enormous datasets spanning tens of thousands of animals — gave him the empirical base for his theoretical work on population structure.

Professor of zoology at the University of Chicago (1926–54). Leon J. Cole Professor of Genetics at the University of Wisconsin-Madison (1955–60); emeritus thereafter but continued publishing for nearly three more decades. Elected to the National Academy of Sciences (1934). National Medal of Science (1966). Balzan Prize (1984). Died 3 March 1988, aged ninety-eight.

The adaptive landscape

Wright introduced the concept in 1932 — a visual metaphor in which the possible gene combinations of a population form a multidimensional space, with fitness mapped as height. Peaks represent locally optimal gene combinations; valleys represent maladaptive combinations. A population under selection will climb toward a peak; the question is how it gets from one peak to a higher one across a valley of reduced fitness.

The metaphor has been enormously influential — and persistently ambiguous. Wright used the landscape metaphor in at least two distinct senses: a genotype-space landscape (where each point is a genotype and the height is its fitness) and a gene-frequency-space landscape (where each point is a population state and the height is mean population fitness). The two behave differently and the conflation has caused confusion in the subsequent literature. The metaphor’s intuitive power has always outrun its formal precision.

Despite the ambiguities, the adaptive landscape gave evolutionary biology a way to think about the structure of the evolutionary problem — the multiplicity of local optima, the role of population structure in navigating between them, and the relationship between selection (which climbs) and drift (which wanders).

Genetic drift

Wright demonstrated that in small populations, random fluctuations in gene frequency — genetic drift — can be a significant evolutionary force alongside natural selection. Drift can fix alleles that are selectively neutral or even mildly deleterious, and can move populations away from local adaptive peaks, potentially allowing them to reach the basins of attraction of higher peaks.

The emphasis on drift put Wright in productive tension with Fisher, who regarded drift as a minor perturbation in the large populations he modelled. The Fisher-Wright disagreement was partly about population size (how small are natural populations, really?) and partly about the relative importance of selection vs drift in driving evolutionary change. The disagreement structured population genetics for decades and was never fully resolved between the two; Kimura’s neutral theory (1968) later extended the argument for drift’s importance to the molecular level.

The shifting balance theory

Wright’s most distinctive and most contested theoretical contribution. The theory proposes a three-phase process:

Phase 1: Drift. In small, partially isolated subpopulations (demes), genetic drift moves populations away from their current adaptive peaks, exploring the fitness landscape in ways that directional selection alone cannot.

Phase 2: Selection within demes. If drift carries a subpopulation into the basin of attraction of a higher peak, selection within that deme drives it up the new peak.

Phase 3: Interdemic selection. The more-fit deme exports migrants to neighbouring demes, shifting them toward the new adaptive combination. The successful combination spreads through the metapopulation.

The theory gives drift a creative role — not as noise that blurs adaptation, but as the mechanism that allows populations to escape local optima and find higher peaks. It requires subdivided population structure, which Wright argued was the typical condition for most species in nature.

The shifting balance theory has been challenged on both theoretical and empirical grounds. Jerry Coyne, Nicholas Barton, and Michael Turelli have argued that the conditions required for the theory to work (small deme size, sufficient isolation, particular fitness landscape topography) are restrictive and rarely met. The empirical evidence for the full three-phase process operating in nature is limited. Defenders argue that the theory identifies a real mechanism even if its relative importance is debated. The theory remains more influential as a framework for thinking about population structure and evolutionary dynamics than as a confirmed description of how most adaptation proceeds.

Path coefficients

Wright developed the method of path analysis (1921) — a statistical technique for partitioning the correlations among variables into direct and indirect causal pathways using a graphical model (path diagrams). The method was developed for his guinea pig genetics work — decomposing the contributions of different genetic and environmental factors to phenotypic variation — but has been widely adopted in sociology, economics, epidemiology, and other fields where causal inference from observational data is needed.

Path analysis is a precursor to modern structural equation modelling. Wright’s insistence that causal diagrams should precede statistical analysis — that the model of causal relationships should be specified before the correlations are interpreted — anticipated the causal inference framework later developed by Judea Pearl.

Where Wright stops

Wright’s programme is populational and structural. It works at the level of gene frequencies in subdivided populations, with the interaction of drift, selection, and migration as the driving forces. Like Fisher’s programme, it abstracts away the organism — the developmental and ecological reality between genotype and fitness is not part of the formal framework. The distinctive limitation is the shifting balance theory’s dependence on specific population-structural conditions (small demes, partial isolation) that may not be typical of the species where most adaptation occurs. Whether population structure plays the creative role Wright assigned it remains the open question his programme leaves behind.

Key works

See also: Darwinism · Fisher · Dobzhansky · Mayr · Kauffman