Artificial selection represents one of humanity’s most transformative scientific and practical endeavors, fundamentally reshaping the global biosphere through deliberate intervention in the reproductive trajectories of living organisms. By selectively propagating individuals exhibiting desirable morphological, physiological, or behavioral traits, human societies have directed evolutionary processes for millennia, creating agricultural staples, domesticated companions, and specialized biomedical models. Understanding artificial selection illuminates not only the foundations of evolutionary theory and quantitative genetics, but also contemporary bioethical challenges concerning genetic diversity, animal welfare, and ecological resilience.
Artificial Selection
1. Concise Definition
Artificial selection, often used interchangeably with selective breeding, is the deliberate, human-mediated process of choosing specific individual organisms to reproduce based on the presence of preferred phenotypic characteristics. Through this non-random reproductive propagation over successive generations, the genetic architecture of the target population shifts, driving evolutionary modification and divergent trait specialization according to human preferences rather than ecological fitness.
Unlike natural evolutionary pathways where differential reproductive success is dictated by ecological pressures and survival fitness within an uncontrolled environment, artificial selection substitutes anthropocentric criteria as the primary selective filter. Humans establish the selective threshold, deciding which genetic variants proliferate, which remain neutral, and which are systematically culled from the breeding pool.
In modern biological sciences, artificial selection bridges classical organismal biology, quantitative genomics, and evolutionary ecology. It operates across empirical scales, ranging from traditional morphological selection practiced by early agrarian societies to modern genomic-assisted selection applied in commercial agribusiness and scientific laboratory research.
2. Etymology & Linguistic Origin
The term artificial selection derives from the Latin artificialis, meaning “belonging to art” or “crafted by human skill” (from ars, meaning art, skill, or craft, combined with facere, meaning to make or do), and selectio, meaning “a choosing out” (from the prefix se-, meaning apart, and legere, meaning to gather, choose, or read). The compound term entered the formal scientific lexicon through British naturalist Charles Darwin in his seminal 1859 treatise, On the Origin of Species.
Darwin purposefully coupled “artificial” with “selection” as an analogical pedagogical device. Recognizing that his contemporaries were intimately familiar with the achievements of Victorian pigeon fanciers, horticulturalists, and livestock breeders, Darwin leveraged the manifest reality of human-directed breeding to render his revolutionary concept of natural selection comprehensible to both scientific peers and the lay public. Over the subsequent century, the term became firmly embedded within evolutionary biology, animal science, and plant genetics.
3. Pronunciation & Grammatical Form
In standard English phonetic notation, artificial selection is pronounced as /ˌɑːrtɪˈfɪʃəl sɪˈlɛkʃən/ in General American phonology and /ˌɑːtɪˈfɪʃl sɪˈlɛkʃn/ in Received Pronunciation. The term functions grammatically as a compound noun phrase, categorized predominantly as an uncountable abstract noun when describing the evolutionary mechanism (e.g., “artificial selection has transformed the wild canid genome”), but occasionally employed as a countable noun when referring to discrete experimental protocols or historical programs (e.g., “multiple distinct artificial selections were performed on the cohort”).
Related morphological and derivative forms include the adverbial phrase artificially selected (participial adjective, as in “an artificially selected cultivar”), the verbal construct to artificially select, and the nominal agent artificial selector. In contemporary quantitative and population genetics literature, researchers frequently use the technical synonym anthropogenic selection or the domain-specific phrase directed selection under domestication.
4. Detailed Conceptual Explanation
At its mechanistic core, artificial selection functions through the manipulation of reproductive differential. In any biological population possessing phenotypic variation that is partly rooted in additive genetic variation (heritability), the non-random allocation of mating opportunities alters the allele frequencies across subsequent generations. When human breeders establish an explicit standard—such as increased grain yield, docility, decorative feather morphology, or drought resistance—and restrict breeding privileges solely to individuals meeting or exceeding this threshold, they enact directional selection upon the underlying quantitative trait loci (QTL).
The mechanics of this process are elegantly captured by classical quantitative genetics through the breeder’s equation, formulated as:
R = h²S
Where R represents the response to selection (the evolutionary change observed in the mean trait value of the offspring generation), h² denotes the narrow-sense heritability (the proportion of total phenotypic variance attributable to additive genetic effects), and S signifies the selection differential (the difference between the mean phenotypic value of the selected parents and the mean phenotypic value of the whole parental generation). Consequently, the pace and trajectory of artificial selection depend directly upon the magnitude of additive genetic variation available and the stringency of the breeder’s culling protocol.
Crucially, artificial selection does not act upon isolated genetic loci in complete independence. Because genomes operate as complex, integrated networks, selecting for a single prominent phenotype frequently induces downstream physiological, developmental, and anatomical alterations through genetic linkage, linkage disequilibrium, and pleiotropy. When a target allele is driven toward fixation by intense selection, physically contiguous chromosomal regions are swept along with it—a phenomenon termed a selective sweep or genetic hitchhiking. This dynamic frequently results in the unintentional fixation of deleterious recessive mutations, explaining why heavily selected domestic breeds often suffer from specific hereditary disorders.
Furthermore, artificial selection manifests in two distinct operational modes: intentional (conscious) selection and unintentional (unconscious) selection. Intentional artificial selection involves an explicit, predefined goal, wherein breeders methodically preserve individuals exhibiting a clear trait, such as rapid growth rates in broiler poultry. Unintentional artificial selection occurs when human harvesting, cultivation, or management practices inadvertently favor particular traits without deliberate foresight. A classic illustration is the historical selection for non-shattering seed heads in early wild cereals; early foragers preferentially collected grains that remained attached to the rachis, gradually favoring non-shattering genotypes without conscious comprehension of Mendelian genetics.
5. Historical Development
The historical trajectory of artificial selection spans over twelve millennia, functioning as the primary biological catalyst for human civilization. Its development can be delineated into four primary historical epochs: the Neolithic Domestication Era, the Empiricist Agricultural Transformation, the Darwinian Synthesis, and the Modern Molecular Revolution.
The origins of artificial selection trace back to the dawn of the Neolithic Revolution, approximately 10,000 to 12,000 years ago in regions such as the Fertile Crescent, the Yangtze River basin, and Mesoamerica. Hunter-gatherer societies transitioning to settled agrarian communities began domesticating wild species, transforming wild wolf progenitors into early domestic dogs (Canis lupus familiaris), teosinte into primitive maize (Zea mays), and wild einkorn wheat into cultivated cereal strains. These early interventions were largely unconscious, driven by natural human preferences for non-aggressive animals, high-yield plants, and non-toxic crop phenotypes.
During the eighteenth century, artificial selection transformed from an intuitive folk craft into a systematic, documented discipline. Pioneering British agriculturalist Robert Bakewell (1725–1795) revolutionized livestock breeding by implementing methodical progeny testing, selective inbreeding (crossing close relatives to “fix” desirable traits), and precise record-keeping. Bakewell successfully established iconic, uniform breeds such as the Dishley Longhorn cattle and New Leicester sheep, demonstrating that biological forms could be rapidly molded through deliberate parental pairings.
The theoretical framework of artificial selection was codified in 1859 by Charles Darwin. In Chapter 1 of On the Origin of Species, titled “Variation under Domestication,” Darwin documented the extraordinary diversity among domestic pigeons (such as Fantails, Pouters, and Tumblers), tracing their lineage back to the wild rock dove (Columba livia). Darwin argued that if human breeders could achieve such vast morphological transformations in mere centuries, nature, operating over geological epochs through natural selection, possessed far greater capacity to generate the entire tree of life. Darwin expanded this thesis extensively in his two-volume 1868 work, The Variation of Animals and Plants under Domestication.
The twentieth and twenty-first centuries witnessed the synthesis of Darwinian selection with Mendelian genetics, led by pioneers of quantitative genetics including Ronald Fisher, Sewall Wright, and Jay Lush (often termed the father of modern animal breeding). The development of statistical methodologies such as the Best Linear Unbiased Prediction (BLUP) by C. R. Henderson in the 1950s enabled breeders to isolate environmental influences from true breeding values. Today, artificial selection has entered the post-genomic era, utilizing high-throughput whole-genome sequencing, marker-assisted selection (MAS), and genomic selection (GS) to calculate genomic estimated breeding values (GEBVs) before phenotypes are physically expressed.
6. Theoretical Foundations
The academic scaffolding supporting artificial selection integrates principles from several foundational disciplines within biology: quantitative genetics, population dynamics, evolutionary synthesis, and developmental canalization.
Central to artificial selection theory is the infinitesimal model, originally formalized by Ronald Fisher in 1918. This framework posits that complex quantitative traits—such as milk volume, adult stature, or crop biomass—are governed by an immense number of polymorphic loci, each exerting a small additive effect. Under this theoretical paradigm, artificial selection does not merely toggle discrete binary genetic switches; rather, it shifts the continuous multivariate distribution of polygenic alleles throughout a population, adjusting phenotypic norms without necessarily exhausting standing genetic variation immediately.
Another foundational cornerstone is the concept of the adaptive landscape or fitness surface, introduced by evolutionary geneticist Sewall Wright. In natural ecosystems, fitness landscapes feature multiple adaptive peaks shaped by environmental constraints. Artificial selection dynamically distorts this landscape, constructing steep, artificial fitness peaks corresponding exclusively to anthropocentric utility. Traits that would represent deep fitness valleys in natural ecosystems—such as the massive musculature of double-muscled cattle or the flightlessness of domestic silk moths—become the supreme adaptive peaks within the human-managed environment, maintained solely by human shelter, feeding, and artificial reproductive management.
The dynamics of artificial selection are also framed by the theory of genetic trade-offs and the principle of energetic allocation. Because an organism possesses a finite metabolic resource budget, intensely selecting for a specific physiological trait (e.g., extreme egg-laying frequency in chickens) inevitably diverts energy away from other physiological domains, such as immunocompetence, skeletal density, or flight capacity. The quantitative architecture of these trade-offs is formalized through genetic correlation matrices (G-matrices), which track how selection pressure applied to Trait A elicits correlated evolutionary shifts in Traits B, C, and D.
7. Key Components, Types & Dimensions
Artificial selection encompasses several distinct methodologies, operational intensities, and biological dimensions:
- Directional Artificial Selection: The most prevalent breeding protocol, wherein breeders continuously favor individuals at one extreme of the phenotypic distribution (e.g., selecting consistently for the highest butterfat content in dairy cattle), driving the population mean steadily in one direction over successive generations.
- Stabilizing (Maintenance) Artificial Selection: Breeding interventions designed to preserve established breed standards and eliminate developmental aberrations. Individuals that deviate from the recognized breed archetype are disqualified from reproduction to maintain uniformity.
- Disruptive (Diversifying) Artificial Selection: The deliberate divergence of a single ancestral population into distinct, specialized breeds. For example, selecting ancestral canids simultaneously for small size and lap-dog companionship versus giant stature and guarding capabilities.
- Mass (Phenotypic) Selection: The classical approach of selecting individuals based solely on their own observed physical traits, without formal documentation of pedigree or ancestry. This approach is highly effective for traits with high narrow-sense heritability.
- Pedigree and Family Selection: A methodological strategy where candidates are chosen based on the performance of their relatives (parents, siblings, or progeny). This is vital for traits with low heritability or sex-limited characteristics, such as selecting dairy bulls based on the lactation metrics of their daughters.
- Marker-Assisted Selection (MAS): An indirect selection process where a target trait is selected based on a linked molecular marker (e.g., single-nucleotide polymorphisms or microsatellites) associated with underlying quantitative trait loci.
- Genomic Selection (GS): An advanced contemporary methodology that utilizes dense, genome-wide molecular markers to estimate the total genetic merit of an individual (Genomic Estimated Breeding Value, or GEBV), bypassing the delay of waiting for full adult phenotypic maturation.
- Unconscious (Inadvertent) Selection: Selection that proceeds without an explicit long-term evolutionary objective, emerging organically from human ecological alterations, harvesting practices, or personal preference patterns.
8. Examples & Illustrative Cases
The evolutionary consequences of artificial selection are illustrated by several hallmark empirical cases across botany, zoology, and experimental biology.
A premier botanical exemplar is the dramatic diversification of the wild mustard plant, Brassica oleracea. Through directional selection focused on distinct anatomical components of this single ancestral species, ancient agriculturalists generated a striking array of common culinary vegetables: selection for enlarged terminal leaf buds produced modern cabbage; selection for shortened, lateral stem nodes created Brussels sprouts; selection for hypertrophied floral stems and immature flower heads produced broccoli and cauliflower; selection for thickened, fleshy stems resulted in kohlrabi; and selection for broad vegetative leaf surfaces yielded kale. Each of these distinct cultivars remains taxonomically classified within the same species, proving the capacity of artificial selection to radically alter plant morphology.
In the animal kingdom, the domestic dog (Canis lupus familiaris) represents an extraordinary demonstration of morphological and behavioral radiation driven by artificial selection. Originating from an extinct Pleistocene wolf ancestor, domestic dogs now exhibit greater morphological variance than any other terrestrial mammalian species—ranging in body mass from the two-kilogram Chihuahua to the eighty-kilogram English Mastiff. Breed diversification has systematically altered skull anatomy (brachycephalic vs. dolichocephalic craniums), limb proportions, coat textures, and instinctive behavioral repertoires (e.g., herding in Border Collies, pointing in German Shorthaired Pointers, and earth-burrowing in Dachshunds).
An illustrative experimental case is the famous Farm-Fox Experiment initiated in 1959 by Soviet geneticist Dmitry Belyaev and continued by Lyudmila Trut at the Institute of Cytology and Genetics in Novosibirsk, Russia. Belyaev sought to uncover the physiological and genetic mechanisms underlying domestication by subjecting captive silver foxes (Vulpes vulpes) strictly to a single selective criterion: tameness toward humans (absence of fear and aggression). Within fewer than ten generations, the selected foxes exhibited affectionate, dog-like social behaviors toward researchers. Astonishingly, this behavioral selection provoked an array of unselected morphological and physiological changes—floppy ears, shortened or rolled tails, depigmented white spotting on the fur (piebald coats), and prolonged developmental windows—an aggregate phenomenon now recognized as the domestication syndrome, linked to altered neural crest cell migration during embryogenesis.
9. Measurement & Assessment
Assessing artificial selection involves evaluating phenotypic shifts, genetic variance, and genomic alterations across generational intervals. Quantitative geneticists and population biologists utilize a rigorous battery of mathematical metrics, statistical models, and molecular tools to quantify selection dynamics.
The foundational metric is the selection intensity (denoted as i), which standardizes the selection differential relative to the phenotypic standard deviation of the trait: i = S / σ_p. Standardizing the selection differential allows researchers to compare the severity of selective culling across disparate populations, traits, and species. Additionally, the realized heritability (h²_realized) is estimated experimentally by calculating the ratio of the realized phenotypic response to the applied selection differential: h²_realized = R / S.
At the molecular level, artificial selection is measured through population genomics tools that identify signatures of historical and contemporary selective sweeps. Key genomic diagnostics include:
- Wright’s Fixation Index (F_ST): Measures the degree of genetic differentiation between the domestic or selected population and its ancestral or unselected wild counterpart. Elevated locus-specific F_ST values indicate genomic regions subject to intense directional selection.
- Nucleotide Diversity (π): Quantifies the average degree of genetic polymorphism within a given chromosomal window. A dramatic drop in nucleotide diversity across a population signifies an artificial selective sweep that purged alternate alleles.
- Cross-Population Extended Haplotype Homozygosity (XP-EHH): Detects high-frequency haplotypes that have swept rapidly through a selected lineage before recombination could break down ancestral linkage disequilibria.
- Genome-Wide Association Studies (GWAS): Identifies the quantitative trait loci that correlate with selected phenotypic variations, clarifying the molecular targets of historical breeding efforts.
10. Applications & Practical Significance
Artificial selection is the technological engine underpinning global agriculture, animal husbandry, industrial biotechnology, and scientific biomedical research.
In global agronomy, artificial selection has driven staggering improvements in agricultural productivity, directly safeguarding international food security. The twentieth-century Green Revolution, spearheaded by Nobel Peace Prize laureate Norman Borlaug, centered on the artificial selection and breeding of semi-dwarf, lodging-resistant, high-yielding wheat and rice strains. Modern crop breeding continues to leverage genomic selection to engineer cultivars exhibiting drought resistance, saline soil tolerance, pest immunity, and enriched micronutrient profiles (biofortification), which are indispensable adaptations in the face of anthropogenic climate change.
In commercial livestock operations, selective breeding has optimized animal productivity to unprecedented levels. In dairy cattle (Bos taurus), the introduction of artificial insemination combined with BLUP-based sire selection elevated average annual milk production per cow from roughly 4,000 kilograms in the mid-twentieth century to over 10,000 kilograms today. Similarly, broiler poultry have been selected for exceptional feed conversion efficiency and rapid breast muscle hypertrophy, vastly reducing the caloric inputs required to produce dietary protein.
Beyond commercial agriculture, artificial selection serves as an indispensable tool in experimental biological research. Laboratories rely upon systematically selected model organisms, such as specialized strains of mice (Mus musculus) and fruit flies (Drosophila melanogaster), to investigate human disease pathogenesis. For instance, selective breeding programs have produced spontaneously hypertensive rats, alcohol-preferring rodent lineages, and tumor-prone animal models, enabling researchers to interrogate complex polygenic human conditions in standardized, controlled settings.
11. Research & Empirical Evidence
A vast body of empirical literature validates the operational principles and theoretical boundaries of artificial selection. Among the most historically significant empirical inquiries is the Illinois Long-Term Selection Experiment on Corn, initiated in 1896 by Cyril G. Hopkins at the University of Illinois and sustained continuously for well over a century.
Hopkins began with 163 ears of an open-pollinated maize variety, establishing four divergent selection regimes: high oil content, low oil content, high protein content, and low protein content. After more than one hundred generations of persistent directional selection, the lineages displayed profound divergence with zero overlap in their phenotypic distributions. The high-oil strain climbed from an initial 4.7% oil content to over 20%, while the low-oil line dropped to nearly zero. This multi-generational study provided empirical evidence that intense directional selection can continually extract phenotypic responses over evolutionary timescales by reorganizing polygenic architectures, challenging the historic notion that selection would instantly stall upon reaching standing genetic limits.
In evolutionary microbiology, long-term selective regimes have provided real-time documentation of evolutionary adaptation. The Long-Term Experimental Evolution project led by Richard Lenski at Michigan State University, tracking more than 75,000 generations of Escherichia coli under defined selective media since 1988, has demonstrated how controlled selection environments drive repeated adaptive peaks, including the revolutionary evolution of citrate-utilizing metabolic capacity under aerobic conditions.
In modern mammalian genomics, high-resolution resequencing studies—such as those published by Andersson and colleagues on domestic swine, chickens, and horses—have pinpointed the precise genetic loci targeted by early human breeders. These investigations revealed that artificial selection predominantly targets genes involved in endocrine regulation, metabolic pathways, neurodevelopment, and embryonic neural crest cell migration, corroborating Belyaev’s physiological hypotheses with molecular evidence.
12. Cultural & Cross-Cultural Considerations
Artificial selection is not merely a biological phenomenon; it is an inherently cultural enterprise. The phenotypic traits prioritized by human societies reflect localized aesthetic paradigms, religious values, socio-economic necessities, and culinary traditions across historical geographies.
For instance, livestock selection among the pastoralist Maasai and Borana communities of East Africa historically emphasized disease tolerance, drought resilience, and specific coat coloration patterns linked to cultural identity and social prestige. In contrast, post-industrial Western livestock breeding focused on standardized economic yield, anatomical homogeneity, and rapid carcass maturation, discarding the phenotypic and cultural diversity intrinsic to traditional pastoral breeds.
Aesthetic preferences have exerted immense cultural selection pressure in horticulture and companion animal breeding. The centuries-long breeding of Japanese Koi carp (Cyprinus rubrofuscus) generated dozens of meticulously classified varieties (such as Kohaku and Showa) evaluated against strict, culturally embedded visual standards of symmetry, color saturation, and grace of movement. Similarly, the dramatic divergence of canine breeds during the Victorian era in the United Kingdom was intimately entwined with socio-economic status; owning specialized lapdogs, hunting hounds, or coursing breeds served as a tangible display of aristocratic leisure and wealth, codifying arbitrary conformational standards into formal kennel club registries.
13. Criticisms, Debates & Limitations
Despite its historic achievements, artificial selection is accompanied by substantial scientific debates, ecological risks, and serious ethical controversies.
The primary biological limitation of artificial selection is the inexorable erosion of genetic diversity within selected populations. Intense directional selection coupled with the widespread utilization of a small number of elite reproductive sires (e.g., through artificial insemination in livestock) inevitably precipitates a severe genetic bottleneck. This escalates the average inbreeding coefficient, culminating in inbreeding depression, reduced reproductive fecundity, and the fixation of deleterious recessive mutations. In companion animals, this dynamic has produced widespread structural welfare issues, most notoriously in brachycephalic dog breeds (e.g., French Bulldogs and Pugs) whose artificially selected flattened facial conformations cause severe, lifelong Brachycephalic Obstructive Airway Syndrome (BOAS).
In industrial agriculture, reliance on hyper-selected, genetically uniform monocultures creates systemic ecological vulnerabilities. When massive agricultural acreage is planted with genetically homogenous crop cultivars, the landscape becomes vulnerable to catastrophic phytopathological outbreaks. A historical warning occurred during the Irish Potato Famine of the 1840s, where reliance on the clonally propagated ‘Lumper’ potato variety left crops defenseless against the late blight pathogen, Phytophthora infestans. A modern equivalent persists in commercial banana production, which relies almost exclusively on the Cavendish monoculture, currently threatened by the fungal pathogen Fusarium oxysporum tropical race 4 (TR4).
Finally, intense philosophical and ethical debates surround the anthropocentric instrumentalization of non-human life. Critics within animal welfare ethics argue that breeding agricultural animals for extreme production metrics—such as broiler chickens whose rapid muscular hypertrophy outpaces skeletal bone mineralization, causing chronic lameness—constitutes structural biological harm. In response, modern livestock genetics has begun integrating “functional traits,” shifting selective criteria away from pure yield toward disease resistance, longevity, and maternal welfare.
14. Related Terms & Distinctions
To preserve analytical clarity, artificial selection must be distinguished from several adjacent biological and biotechnological concepts:
- Natural Selection: Evolutionary change driven by differences in reproductive success resulting from environmental interactions, ecological competition, predation, and abiotic stressors, without deliberate human intervention. Artificial selection is an anthropocentrically guided subset of the broader principle of selection.
- Sexual Selection: A distinct mode of natural selection where evolutionary traits are driven by intrasexual competition for mates or intersexual mate choice. While mate choice in sexual selection is dictated by conspecifics within the species, artificial selection usurps this dynamic by having humans dictate mating access.
- Genetic Engineering (Recombinant DNA Technology): The direct molecular manipulation, insertion, deletion, or editing of an organism’s DNA using technologies such as CRISPR/Cas9 or transgenesis. Unlike classical artificial selection, which relies on selecting whole organisms possessing naturally occurring or induced mutations over generations, genetic engineering modifies the genome directly in a single generational step.
- Domestication: The broader co-evolutionary, sociocultural, and ecological process whereby an entire species adapts to humans and the captive environment. Artificial selection is the primary mechanistic tool through which domestication is biologically achieved, but domestication also encompasses behavioral habituation, ecological niche construction, and mutualism.
- Eugenics: A discredited, unscientific, and deeply unethical social and political ideology that sought to apply principles of artificial selection to human populations through state-mandated reproductive control, forced sterilization, and racialized discrimination. Mainstream scientific and bioethical consensus rejects eugenics, recognizing human biological diversity and autonomy as fundamental ethical absolutes.
15. Summary & Key Takeaways
Artificial selection stands as one of the most powerful empirical validations of evolutionary theory, illustrating how non-random reproductive propagation can swiftly reshape biological morphologies, life histories, and genomic landscapes. By substituting human utility, aesthetics, and curiosity for wild ecological fitness, selective breeding has orchestrated the domestication of plants and animals, providing the energetic foundation for human civilization. Yet, as modern genomics accelerates our capacity to edit and curate genomes, the management of artificial selection demands vigilance: balancing short-term phenotypic gains against the enduring biological imperative of genetic diversity, organismal welfare, and systemic ecological health.
References
- Bakewell, R. (1790). Observations on the Influence of Soil and Climate upon Wool and Sheep. London: Agricultural Society.
- Belyaev, D. K. (1979). Destabilizing selection as a factor in domestication. Journal of Heredity, 70(5), 301–308. https://doi.org/10.1093/oxfordjournals.jhered.a109263
- Darwin, C. (1859). On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life. John Murray.
- Darwin, C. (1868). The Variation of Animals and Plants under Domestication. John Murray.
- Falconer, D. S., & Mackay, T. F. C. (1996). Introduction to Quantitative Genetics (4th ed.). Longman.
- Lush, J. L. (1945). Animal Breeding Plans (3rd ed.). Iowa State College Press.
- Trut, L., Oskina, I., & Kharlamova, A. (2009). Animal evolution during domestication: the domesticated fox as a model. BioEssays, 31(3), 349–360. https://doi.org/10.1002/bies.200800070
- Wright, S. (1968). Evolution and the Genetics of Populations: Vol. 1, Genetic and Biometric Foundations. University of Chicago Press.