Chapter VI. Difficulties Of The Theory
Charles Darwin · 1872. Read the original ↗
Darwin takes the strongest objections to descent with modification by natural selection (missing intermediates, bats, the eye, electric organs, useless or beautiful structures) and argues that none is fatal. His case rests on graded series among living forms, the extinction of intermediates, changes of function, and explicit statements of what would refute the theory.
The chapter builds on real observations (graded structures, variation, homology, correlations of form and function) and states its refuting cases openly, which is rare and creditable. But its central step, that selection acting on slight variations built complex organs and new structures, is inferred from living series, not observed or tested. It is extended by deep time in the eye and bat arguments, so those parts grade lower. Hedged phrasing like 'no insuperable difficulty' keeps most claims from being tested.
Claims, graded
- C1InferredC
Transitional varieties are rarely found, either living or fossil, because intermediate forms are numerically scarcer, are supplanted by the better-adapted forms on either side, and are lost from an imperfect geological record.
“the very process of natural selection constantly tends, as has been so often remarked, to exterminate the parent forms and the intermediate links.”
Evidence offered Reports (Balanus, Watson, Gray, Wollaston) that intermediate varieties are numerically rarer than the forms they link, abrupt range edges, and a sheep-breeding illustration. The fossil record is deferred to a later chapter.
- 1Observation//The rarity of intermediate varieties and the sharp range edges are observed, though mostly from informants' reports.
- 2Hypothesis//It is a definite mechanism, but several overlapping explanations are offered, so any one gap can be covered by another.
- 3Prediction/Rare intermediates are fitted after the fact to the absence of links, and only partly predicted.
- 4Experiment—No controlled test is given; the sheep scenario is an illustration, not a test.
- 5Replication/Several naturalists report similar rarity, but no independent systematic survey is cited.
- 6Falsifiability/Independent creation is mentioned, but the absence of links is explained by the theory however the evidence falls.
What would move it up Systematic surveys of contact zones showing intermediate forms with lower numbers and shorter persistence, plus fossil sequences with dense sampling showing intermediates where preservation is good.
- C2InferredC
A complex eye can be accounted for by natural selection: graded eyes exist among living animals, eyes vary and are inherited, and each grade serves its owner, so there is no logical impossibility in the eye arising through successive slight changes.
“if numerous gradations from a simple and imperfect eye to one complex and perfect can be shown to exist, each grade being useful to its possessor, as is certainly the case”
Evidence offered A range of living eye types from pigment spots to compound and vertebrate eyes, the embryonic origin of the human lens, and Wallace's argument that lens improvements need not be simultaneous.
- 1Observation//Graded eye structures in living species, and variation and inheritance in general, are observed, but not the steps in any lineage.
- 2Hypothesis//The claim of gradual, useful steps could fail if a step were shown to be useless or impossible.
- 3Prediction/Little is predicted in advance beyond the expectation that further graded forms would be found.
- 4Experiment—No test is offered that selection builds an eye or any comparable organ.
- 5Replication/Several authorities (Müller, Owen, Jourdain, Hensen) describe the series, but nothing is independently reproduced.
- 6Falsifiability//Darwin states a refuting case and answers the 'simultaneous change' objection, though the origin of the first light-sensitive nerve is set aside.
What would move it up Evidence that the specific variants needed at each step (lens, retina, focusing) actually arise, are heritable and raise fitness, for example from developmental genetics or fossil or phylogenetic series of a lineage. Later molecular work on eye development bears on this.
- C3ExtrapolatedD
Given millions of years, millions of individuals and unfailing preservation of slight improvements, a living optical instrument superior to glass could be produced by variation, multiplication and natural selection.
“Let this process go on for millions of years; and during each year on millions of individuals of many kinds”
Evidence offered A thought-experiment with a hypothetical transparent layer that is always slowly changing, and a selecting 'power' that preserves each improvement; support rests on the graded eyes and on variation under domestication argued elsewhere.
- 1Observation/Slight variation and selection of small differences are observed, but not accumulation to a new optical organ.
- 2Hypothesis/Stated as 'may we not believe', so it is permissive and hard to fail.
- 3Prediction—No risky prediction accompanies the thought-experiment.
- 4Experiment—Nothing was tested; it is a thought-experiment.
- 5Replication—No independent reproduction is possible or cited.
- 6Falsifiability/The scenario is framed so that deep time absorbs any shortfall, though the chapter elsewhere gives a refuting case.
What would move it up Observed or experimental accumulation of several coordinated, individually beneficial changes in an optical or other composite structure, or reconstruction of the steps in a documented lineage with rates shown sufficient.
- C4ExtrapolatedD
A bat could have arisen from a non-flying insectivore by way of gliding forms like flying squirrels and the 'flying lemur', each grade being useful and preserved by selection.
“Nor can I see any insuperable difficulty in further believing it possible that the membrane-connected fingers and fore-arm of the Galeopithecus might have been greatly lengthened by natural selection”
Evidence offered A graded series of squirrels up to gliders, the gliding Galeopithecus, and bats with wing membranes reaching to the tail; Darwin admits that he can give only a few examples and that no gliding links connect Galeopithecus to other insectivores.
- 1Observation/Gliding mammals are observed, but no sequence leading to powered flight is.
- 2Hypothesis/It is posed only as 'possible', which is loose enough to resist testing.
- 3Prediction—No prediction about intermediates is stated that could fail.
- 4Experiment—No test of any kind is offered.
- 5Replication—Nothing is independently reproduced; the chain is a reasoned scenario.
- 6Falsifiability/The limited claim of 'not impossible' is hard to refute, though Darwin says it is not to be read as the actual route.
What would move it up Fossil or developmental evidence of gliding forms in the actual line of ancestry of bats, with their changes in structure. Later fossil bats are already fully winged, and colugos are no longer placed near bats, so the specific proposal is not supported.
- C5InferredC
Organs can change function in the course of descent, as in the swim-bladder becoming a lung or the ovigerous frena of barnacles becoming gill-like branchiae. Homology shows the conversion and removes the objection that intermediate stages would be useless.
“hence there is no reason to doubt that the swim-bladder has actually been converted into lungs, or an organ used exclusively for respiration.”
Evidence offered The homology of swim-bladder and lungs, fishes that breathe both with gills and swim-bladder, organs doing two functions at once (dragonfly larva gut, Hydra), and barnacle frena and branchiae that 'graduate into each other'.
- 1Observation//The homologies and organs doing more than one function are observed, but the conversion itself is not.
- 2Hypothesis//It makes definite claims about homology and the direction of change, which could be wrong.
- 3Prediction/The cirripede and swim-bladder cases are explained after the fact, with little risk to the claim.
- 4Experiment—No controlled test of functional conversion is described.
- 5Replication/Several naturalists agree on the homologies, but the conversion is not independently tested.
- 6Falsifiability/Rival explanations such as separate creation are dismissed in a sentence rather than compared on shared predictions.
What would move it up Phylogenetic polarity tests and developmental genetics showing the direction of change and intermediate states of function. Later work suggests lungs are ancestral and swim-bladders derived, reversing Darwin's direction while keeping the homology, and the tracheal origin of insect wings is now disputed.
- C6InferredC
Natural selection acts only for the good of each being, so no structure can be formed solely for another species' benefit, for beauty to humans, or to injure its owner. Structures not useful now are explained by inheritance, laws of growth or sexual selection. Selection yields only relative, not absolute, perfection.
“If it could be proved that any part of the structure of any one species had been formed for the exclusive good of another species, it would annihilate my theory”
Evidence offered The rule that wind-pollinated flowers lack showy corollas, colour in fruit linked to seed dispersal, imperfections such as the human eye and the bee sting, and cautionary examples (vulture, turkey, skull sutures) where assumed adaptation was doubted.
- 1Observation//The flower and fruit correlations and the imperfections of various structures are observed.
- 2Hypothesis///The claim is sharp: a single structure shown to serve only another species would refute it.
- 3Prediction//The rule that colour goes with insect pollination is a real prediction checked on many species, though not framed in advance.
- 4Experiment—No experiment measures the fitness effects of the structures discussed.
- 5Replication/Darwin says he has found no exception, but no independent surveys are cited.
- 6Falsifiability//Refuting cases are stated openly, though costly traits are rescued by appeal to community benefit.
What would move it up Fitness measurements of variants differing in the trait, such as stinging versus non-stinging colonies or colourless versus coloured flowers. A search for a documented structure that benefits only another species would test the claim directly.
Reasoning check
- PossibleClaim C6
“if on the whole the power of stinging be useful to the social community, it will fulfil all the requirements of natural selection, though it may cause the death of some few members.”
Where benefit is measured separately, as in flower colour and insect visits or cattle and flies, 'useful' is independent of survival. For the fatal sting, the drones and the queen's killing of her daughters, benefit to the community is mostly inferred from the trait existing. Colony-level fitness measurements of stinging versus non-stinging workers would settle whether 'for the good of the community' is an independent claim.
- PossibleClaim C3
“Let this process go on for millions of years; and during each year on millions of individuals of many kinds”
Darwin offers real evidence of graded eyes among living animals and of observed variation and selection under domestication. He has not shown selection building a coordinated organ of this kind, and the length of time stands in for evidence that the steps are available. Observed or reconstructed accumulation of coordinated beneficial changes in a documented lineage would settle it.
- PossibleClaim C4
“Although no graduated links of structure, fitted for gliding through the air, now connect the Galeopithecus with the other Insectivora, yet there is no difficulty in supposing that such links formerly existed”
The scenario rests on a real graded series of squirrels and on gliding membranes in living mammals, but the links for the actual lineage are assumed. Darwin does say these are possible routes, not the actual one. Fossil or developmental evidence of gliding ancestors of bats would test it.
- Clear caseThe argument as a whole
“Why, on the theory of Creation, should there be so much variety and so little real novelty?”
The argument that graded, linked structures favour selection over separate creation assumes that a Creator would not produce discrete novelty or sudden leaps. That premise about what a Creator would do is neither stated openly nor argued. The observed pattern of homology and gradation is real evidence, but the contrast with creation depends on this assumption.
What it does well
- Darwin states what would refute his theory: a complex organ that could not have been formed by successive slight modifications, or a structure formed solely for another species' good.
- He treats the difficulties candidly, calling the eye 'absurd' at first sight, admitting that the electric organs are a case of special difficulty, and conceding that he cannot name the steps for them.
- He gives observed cases of intermediate habits and structures, such as Mustela vison, the squirrels, the La Plata woodpecker and the diving petrel.
- He warns against assuming that every trait is an adaptation, citing growth laws, correlation, sexual selection and inheritance, and the vulture, turkey and skull-suture cases.
- Müller's test on air-breathing crustaceans is a real comparison of rival explanations: descent from a common ancestor predicts the shared features but different air-breathing organs, while functional convergence predicts the differences.
- He limits what selection can do, saying it produces only relative perfection, and shows this with imperfections such as the bee's sting and the human eye.
- He disowns the telescope analogy in the eye argument, asking 'may not this inference be presumptuous?'
Limits of this evaluation This is one section of a longer work, and several supports are deferred to other chapters: the imperfection of the geological record, variation under domestication, instinct and hybridism. Those were not credited or counted against the claims here. It is a nineteenth-century work, so later findings (genetics, fossils, evo-devo) are noted only where they bear directly. On the focus: the agency language ('picks out with unerring skill', 'intently watching', 'for the good of each') is introduced as 'represented by' or shorthand, and the argument survives restating it as differential survival of variants, so I did not diagnose reification, though 'unerring' contradicts Darwin's own admission of imperfections. The breeder analogy was not diagnosed as a false analogy: the sheep example depends on population size supplying variants, not on foresight, and the eye-telescope analogy is explicitly rejected. Filter-as-source was not diagnosed because variation is named as the source ('natural selection can do nothing until favourable individual differences or variations occur'), though the origin of the first light-sensitive nerve is set aside and Darwin admits ignorance of the causes of variation. Tautology applies only where benefit is inferred from existence. Burden of proof: Darwin's 'I can find out no such case' and 'it would be extremely bold to maintain' are fair as a defence against refutation but are not evidence for his own claims.