Chapter I. Variation Under Domestication
Charles Darwin · 1872. Read the original ↗
Domestic breeds arise because organisms vary, much of the variation is heritable, and breeders (methodically or unconsciously) accumulate it over generations. The chapter argues, with domestic pigeons as the worked example, that this is a guide to how natural species may have arisen.
The core results are well observed and widely replicated: heritable variation exists and breeders accumulate it into large changes, and Darwin's pigeon argument is a careful, tested inference. The weaker steps are the causal accounts of where variation comes from, including the inheritance of use effects, and the passage from domestic divergence to the origin of species in nature, which this chapter largely defers to later chapters. Claims about breeders are graded higher than those about nature.
Claims, graded
- C1ObservedB
Selective breeding of heritable variation, by choosing the best individuals generation after generation, produces large and cumulative changes in form and performance, within a human lifetime and with no need for crossing distinct breeds.
“It is certain that several of our eminent breeders have, even within a single lifetime, modified to a large extent their breeds of cattle and sheep.”
Evidence offered Breeders' reports and records (Bakewell, Youatt, Somerville), the steady increase in gooseberry size, improved florists' flowers, and the dozens of pigeon breeds. Darwin also notes that the best breeders avoid crossing and that results depend on the number of individuals kept and on close attention.
- 1Observation///Rapid change under selection in livestock and plants is directly documented in records, animals and plants.
- 2Hypothesis//The claim is clear about direction and scale, but it is stated qualitatively, without measured rates or heritabilities.
- 3Prediction//Darwin notes that selection should fail with few individuals, no isolation or little attention, and the cases he cites bear this out.
- 4Experiment//Breeding programmes are in effect controlled trials, but the chapter reports them anecdotally, without controls or measurements.
- 5Replication///Many breeders, species and centuries independently reproduce the result.
- 6Falsifiability//Crossing and direct effects of conditions are considered as rivals and judged insufficient, though the unknown mechanism of heredity limits how far they are ruled out.
What would move it up Replicated, measured selection lines with unselected controls and recorded heritabilities, which later quantitative breeding experiments supplied.
- C2InferredB
All the domestic pigeon breeds, however different, descend from the wild rock-pigeon, Columba livia, rather than from seven or eight separate wild species.
“we may safely conclude that all our domestic breeds are descended from the rock-pigeon or Columba livia with its geographical sub-species.”
Evidence offered Several lines taken together: no suitable wild ancestors are known, and none have become feral. Crossed breeds, none of them blue, throw back blue birds with the rock-pigeon's wing-bars and tail marks. All breeds are interfertile. Darwin made these crosses himself.
- 1Observation//The crosses, the colour reversion and the fertility were observed directly, but ancestry in the distant past was not.
- 2Hypothesis///The claim names a single ancestor and states what would count against it.
- 3Prediction//Reversion to rock-pigeon colours is what common ancestry expects, but Darwin explains it after seeing it rather than predicting it.
- 4Experiment//His own controlled crosses (fantail x barb, barb x spot) test a specific expectation.
- 5Replication//He cites several crosses and other breeders' observations, though not independent trials.
- 6Falsifiability///The rival hypotheses of multiple stocks, extinct stocks and recent crossing with the rock-pigeon are each stated and answered with reasons.
What would move it up Independent comparison of breed genomes with wild rock-pigeons; later genomic work has done this and agrees with Darwin's conclusion.
- C3InferredC
Unconscious selection, in which people simply breed from their best animals and plants without aiming to change the breed, would over centuries substantially modify breeds. It also explains why the history of most breeds is unrecorded and why wild ancestors of old crops are hard to identify.
“we may infer that this process, continued during centuries, would improve and modify any breed”
Evidence offered Examples such as King Charles spaniels, pointers, race-horses versus Arabs, and heavier cattle. Youatt's report that two flocks of Bakewell's Leicester sheep, bred pure for fifty years, now look like different varieties. The improvement of pears and heartsease since classical times.
- 1Observation//The breed changes are documented, but the claim that unconscious selection caused them is not observed directly.
- 2Hypothesis//The claim is reasonably specific, but a gradual and unrecorded process can fit almost any outcome.
- 3Prediction/Little is predicted in advance, and most examples are explained after the fact.
- 4Experiment/No controlled trial separates unconscious selection from drift, crossing or environment.
- 5Replication/Many cases are cited, but each is a historical reconstruction rather than a repeat of a test.
- 6Falsifiability/Darwin notes that crossing with the foxhound mainly changed the pointer, but he does not rule out chance divergence between flocks.
What would move it up Replicate lines founded from one stock, under recorded selection regimes with unselected controls, showing that divergence tracks the selection actually applied rather than chance or environment.
- C4ExtrapolatedD
Domestic breeds differ from one another as much as, or more than, closely related natural species do, and there is no sharp line between variety and species. So change under domestication gives grounds to take seriously that natural species descend from other species.
“may they not learn a lesson of caution, when they deride the idea of species in a state of nature being lineal descendants of other species?”
Evidence offered A score of pigeons that an ornithologist would rank as distinct species, and breeds placed in different genera. Domestic races whose origin from one or several species experts dispute. Domestic varieties that differ in size, colour and fleetness more than natural species in a genus. Darwin defers an account of how genera arise in nature to later chapters.
- 1Observation//Large morphological divergence within domestic species is observed, but the origin of reproductively isolated natural species is not.
- 2Hypothesis//The comparison is clear, but it depends on subjective taxonomic ranking and on shifting definitions of species.
- 3Prediction/No prediction is made about natural species within this chapter.
- 4Experiment/Selection experiments in domesticates are shown, but none tests whether natural populations behave the same way.
- 5Replication/Domestic divergence recurs across species, but replication of the extrapolated step is deferred.
- 6Falsifiability//Darwin states a test: strong reversion under constant conditions would show that domestic races teach nothing about species.
What would move it up Observed cases in wild populations where selection produces heritable divergence, with reproductive isolation, and a showing that the isolation, population size and persistence that breeders supply also occur in nature.
- C5InferredC
Changed conditions of life are the main cause of variability in domestic plants and animals, acting directly on the body and indirectly through the reproductive system. The nature of the organism matters more than the nature of the conditions.
“Changed conditions of life are of the highest importance in causing variability, both by acting directly on the organisation, and indirectly by affecting the reproductive system.”
Evidence offered Domesticated stocks vary more than wild ones. The reproductive system is sensitive to confinement. Bud 'sports' appear in uniform conditions and similar variations in different conditions. Darwin admits that the laws of variation and inheritance are largely unknown.
- 1Observation//Greater variability under cultivation, sporting buds and breeding failures in captivity are observed.
- 2Hypothesis/'Conditions of life' is vague, and Darwin concedes that the laws involved are unknown.
- 3Prediction/The hypothesis predicts little beyond variability that already exists.
- 4Experiment/A few observations, such as water affecting seed set, are experimental in kind, but there is no systematic test of the cause.
- 5Replication/Many horticulturists report variation, but not the specific causal claim.
- 6Falsifiability//Darwin notes that similar variations arise under unlike conditions and different ones under like conditions, which counts against simple environmental causation.
What would move it up Controlled rearing of genetically identical stock under defined environmental treatments, measuring new heritable variants, and sequencing to identify the changes. Later mutation research shows that much variation arises from DNA changes unrelated to the conditions that favour them.
- C6InferredC
Increased use or disuse of parts produces inherited changes, as seen in the lighter wing bones and heavier leg bones of domestic ducks and in the larger udders of milked cattle.
“this change may be safely attributed to the domestic duck flying much less, and walking more, than its wild parents.”
Evidence offered Measured bone weights of domestic versus wild ducks. Udder development in cows and goats where they are milked. The suggestion that ears droop because muscles are not used.
- 1Observation//The differences in bone proportion and udder size are measured or recorded.
- 2Hypothesis//The claim is specific, but its hedges ('probably', 'seems probable') leave it elastic.
- 3Prediction—The comparison was made after the differences were seen, and nothing was predicted in advance.
- 4Experiment—No rearing experiment separates inherited effects from breeder selection, correlated variation or plasticity.
- 5Replication—No independent test of the inheritance of use effects is offered.
- 6Falsifiability/Rival explanations (selection by breeders, relaxed selection, non-inherited plasticity) are not addressed.
What would move it up Raise wild and domestic ducklings under identical activity regimes across generations and measure bone proportions, with controls for selection. Later genetics has largely cut against inheritance of acquired effects in this form.
Reasoning check
- Clear caseClaim C6
“this change may be safely attributed to the domestic duck flying much less, and walking more, than its wild parents.”
Lighter wing bones would follow from inherited effects of disuse, but they would also follow from breeders' selection, relaxed selection on flight, correlated variation or non-inherited plasticity. The text offers the measured difference as evidence but supplies no test that separates these rivals. Raising wild and domestic ducklings together under the same activity would settle it.
- PossibleClaim C3
“the difference between the sheep possessed by these two gentlemen is so great that they have the appearance of being quite different varieties.”
Darwin offers this report from Youatt (two flocks bred pure from one stock for fifty years) as an illustration of unconscious selection. The evidence is real, and breeders do select within flocks. But sampling chance in small closed flocks, differences in feed or environment, and undetected crossing would also produce divergence. Records of what each breeder actually chose, or replicated lines with controls, would settle it.
- PossibleClaim C4
“may they not learn a lesson of caution, when they deride the idea of species in a state of nature being lineal descendants of other species?”
The text offers real evidence for the narrower step: pigeon breeds diverge morphologically beyond what ornithologists would rank as species, and varieties and species are hard to separate. The step to the origin of natural species rests on the promise that how genera arise in nature will be explained later. It is not shown that nature supplies the large stocks, isolation and sustained attention that Darwin says make breeders succeed. Observed selection producing reproductive isolation in wild populations would settle it.
What it does well
- Separates what is observed (breeder results, pigeon diversity, inheritance records) from what is inferred, and repeatedly marks his own uncertainty (for example 'the laws governing inheritance are for the most part unknown').
- The pigeon-ancestry argument weighs several rival hypotheses (multiple stocks, extinct stocks, recent crossing) with reasons, and includes his own crossing experiments and the reversion to rock-pigeon colours.
- Names conditions that would make his claims fail: no selection effect with small numbers, strong reversion under constant conditions, and the limit on how far traits can be pushed.
- Is candid that human selection works only on variation 'first given to him in some slight degree by nature', so variation and selection are kept distinct. Breeders' criteria (size, colour, speed, shape) are defined independently of survival.
- Notes the circumstances that make human selection effective (large numbers, strict attention, preventing crosses), which shows where the breeder analogy rests on human conditions.
Limits of this evaluation This is one chapter of seventeen and is graded as a historical work: for what Darwin knew and argued in 1872, not for later genetics. The mechanism of heredity was unknown to him, and later work (mutation, Mendelian and molecular genetics, pigeon genomics) supports the pigeon ancestry and the effectiveness of selection but undermines the inheritance of use effects. On the requested focus on the language of agency: the selecting agent in this chapter is literally the breeder, so the agent language is not reification. Phrases such as 'nature gives successive variations' are shorthand for variation arising from causes Darwin openly says are unknown, and the argument survives without them. 'Best' and 'improved' are defined by traits breeders measure rather than by survival, so there is no tautology. Selection is credited only with accumulating variation that already exists, so there is no filter-as-source error. The false-analogy question is largely deferred: the chapter's one use of 'natural selection' (differential success of constitutions in domestic animals kept by savages) is a brief hypothesis with the detail promised later, and whether nature supplies the isolation, numbers and persistent direction of breeding is for later chapters.