Observed, or Only Inferred?

Evolution has been watched in small steps within species: a few mutations, or shifts in variants already present. The large gains, such as new genes, fish antifreeze and species radiations, are reconstructed from comparisons. Some famous cases sold as observed were never watched, and several design and young-earth claims, including Axe's protein rarity and genetic entropy, are inferences too.

By AI (Claude Opus 5.5, reasoning effort 15) Draft:  September 2026

"Homoploid hybrid speciation in animals has been inferred frequently from patterns of variation, but few examples have withstood critical scrutiny. Here we report a directly documented example, from its origin to reproductive isolation." — Sangeet Lamichhaney, Peter and Rosemary Grant and colleagues, 20181

Summary

Some changes in living things were seen happening.  Others were worked out afterwards from the shape of a family tree, a molecular clock or a fossil sequence.  Both are science.  They are not the same kind of evidence, and both sides of the origins debate have presented the second kind as if it were the first.  This article sorts the cases.  It uses one test and applies it to evolutionary and creationist claims alike.

The test has four grades:

  • Watched: someone recorded the before and the after, and usually the steps between. The freezer of the Long-Term Evolution Experiment, the Grants' birds on Daphne Major and a dated series of hospital isolates are examples.
  • Bracketed: the change fell between two dated observations, and the steps between were not seen. A lizard population released in 1971 and sampled in 2004 is bracketed.
  • Reconstructed: the change is inferred by comparing living genomes or fossils. No dated observation of the earlier state exists. The result carries the assumptions of its method: a tree, a clock rate, a calibration.
  • Asserted: the claim is stated without the evidence any of the first three grades requires.

The findings:

  1. The watched cases are real and many. Bacteria evolved citrate use in a flask. A virus gained a new receptor in about two weeks. Finch beaks shifted with droughts and shifted back. Moths turned black and pale again. Hospital enzymes gained resistance to new drugs. All of these were seen, and the site concedes the gains among them.
  2. The watched cases are small. Each is one to a handful of mutations, a new copy of something already present, or a shift in how common an existing variant is. They happened within a species, over days to two centuries. Benchmarking Evolution scores them, and Measuring Biological Information gives the scoring rules.
  3. The large gains are reconstructed. The antifreeze genes of polar fish, the de novo genes of humans, the age of the finch beak variants and the cichlid radiations are all inferred from comparisons. Nobody watched any of them. That does not make them false. It means their strength rests on the method that produced them.
  4. Several famous "observed" cases were never observed. The London Underground mosquito did not evolve in the Underground. The nylonase frameshift was a proposal that its readers reported as a finding. The Italian wall lizards' gut valves have not been shown to be inherited.
  5. The design side has the same fault. Douglas Axe's "may be as low as 1 in 1077" became "only 1 in 1077 sequences are functional." William Dembski's flagellum probability priced a process nobody proposes, and he later set it aside. James Crow's conditional 1% to 2% decline per generation in fruit-fly terms became a 6,000-year history of human decline. In each case the originator hedged and the populariser dropped the hedge.
  6. Ministry claims fail the test in both directions. "Mutations only lead to a loss of functional systems" is an assertion that watched cases contradict. Creation Ministries International already tells creationists not to say "there are no beneficial mutations." The narrow claims that survive are observations too: no antibiotic-destroying enzyme has been watched arising from DNA that was not a gene.
  7. The creationist "operational versus origin science" distinction is sound when applied evenly and unsound when used to dismiss. Its authors wrote that special creation "is not falsifiable. But neither is macroevolution." Rapid post-Flood speciation, the boundaries of created kinds and the harmlessness of microbes before the Fall are reconstructions or assertions under that same test.

Verdict.  Of 38 claims graded here, 25 come from the evolutionary side:  13 watched at their core, 4 bracketed, 6 reconstructed and 2 asserted as observed when they were not.  The design and creationist side has 13:  4 rest on observation, none are bracketed, 5 are reconstructed or modelled, and 4 are asserted.  What was watched is small, and what is large was inferred.  That supports the site's argument about extrapolation.  It does not show that the inferences are wrong, and the young-earth alternatives are inferences too.

Part 1: The test

Four grades

The grades turn on one question:  was the earlier state recorded by someone at the time?

A watched change has the earlier state on record and the later state on record, usually with samples in between.  Richard Lenski's group froze samples of each of twelve Escherichia coli populations "every 500 generations, providing a rich “fossil record”", and the samples can be revived.2  Peter and Rosemary Grant measured nearly every medium ground finch on Daphne Major every year from 1973.  A hospital that stores dated isolates of a bacterium has the same kind of record.

A bracketed change has a dated before and a dated after with nothing between.  It shows that a change happened in a known interval.  It does not show how, and it may not show whether the change was inherited.

A reconstructed change has no recorded earlier state.  It is inferred from the pattern among living organisms or among fossils.  A gene shared by some species and absent from their relatives is inferred to have arisen in their common ancestor.  A difference between two sequences divided by a mutation rate gives a date.  The inference is only as good as the tree, the rate and the calibration that went into it.

An asserted change is claimed without the record any of the first three grades needs.  Most asserted cases in this series began as honest reconstructions and lost their hedges as they were retold.

The membership rule of this section of the site follows the grades.  Only watched and bracketed changes count as observed evolution.  Reconstructions go to the New Genes articles, the phylogeny articles and the fossil articles, where their methods can be judged.

Inference is science, and it has a cost

Inference about the past is not second-class science.  The philosopher Carol Cleland set out the case in 2002.  She noted that historical science "is sometimes held to be inferior on the grounds that its hypothesis cannot be tested by controlled laboratory experiments."  Her paper "rejects the claim that historical research is epistemically inferior."3  Her reason is a feature of nature:  one past event leaves many traces, so a later observer can often find a "smoking gun," a trace "that unambiguously discriminates one hypothesis from among a set of currently available hypotheses."4  The asteroid impact at the end of the Cretaceous is her example.  Its iridium layer and shocked quartz were found long after the event, and they settled the question.

Patrick Forber and Eric Griffith, who defend historical science, also state its cost.  They ask how, "with no way to test any claim against the actual event in real time," historical claims can "become more than mere just-so stories."4  Their answer is that strong support "consists in the consilience of multiple independent lines of evidence," and they add that "the general conditions for this are strict."4

So a reconstruction is worth what its independent checks are worth.  A single tree built on a single calibration is weak.  Several methods on independent data that agree are strong.  The weak case is common in practice, and tuberculosis shows how much a calibration can change the answer.  Iñaki Comas and colleagues found "no significant correlation between isolation time and phylogenetic divergence" among dated modern strains, so they "used as initial calibration points several key dates in human evolution" and dated the tuberculosis complex to about 70,000 years.5 5 5  Kirsten Bos and colleagues calibrated instead on radiocarbon-dated ancient genomes and got "less than 6,000 years ago."6  The data differ, but the main difference is the calibration, and the answers differ by more than a factor of ten.  See Mycobacterium tuberculosis for the detail.

The researchers who do the watching draw the same line.  When the Grants' team documented a new finch lineage on Daphne Major, they opened their paper by saying that hybrid speciation in animals "has been inferred frequently from patterns of variation, but few examples have withstood critical scrutiny.  Here we report a directly documented example".1  The distinction in this article is theirs as much as anyone's.

The creationist version of the distinction

Creationist writers have drawn the line for forty years, under the names operational and origin science, or observational and historical science.  Norman Geisler and Kerby Anderson set it out in 1987:  "a science which deals with origin events does not fall within the category of empirical science, which deals with observed regularities in the present.  Rather, it is more like a forensic science, which concentrates on unobserved singularities in the past."7  They were even-handed about what followed:  "In the strict scientific sense, it is true that special creation is not falsifiable.  But neither is macroevolution."  In a footnote they classed microevolution as "an empirical science, since it can be observed in the present".7

Answers in Genesis keeps the same symmetry in its official statements.  Troy Lacey answered a reader in 2011:  "we have stated that neither creationism nor cosmic evolution nor Darwinian biological evolution is observational science, and they are not observable, testable, repeatable, falsifiable events."8  The reader had put the standard criticism in one line:  "Or is “historical” simply anything which contradicts and “observational” anything that supports you?"8

Where the creationist distinction is right.  It is right that an inferred event carries assumptions that a watched event does not.  It is right that a watched case of one mutation does not by itself show that a thousand coordinated changes happened the same way.  And Geisler and Anderson were right that the same limit binds creation claims.

Where it goes wrong.  It goes wrong in three ways.  First, it treats all inference as equally uncertain, when Cleland's point is that some historical inferences are settled by traces as firmly as any experiment.  Second, it is often applied to one side only.  In a 2016 Answers in Genesis article on a disputed fossil snake, Elizabeth Mitchell wrote that "“Origins science” is not observational science" and that "Evolutionary transitions from one kind of animal to another have never been observed."9  That is true, and a created kind or a post-Flood diversification has never been observed either.  Third, it is used to wave away evidence without engaging it.  Reading the pattern of shared mutations across species is a legitimate test.  The right answer is to assess the method, as Phylogenetic Congruence does, not to rule the whole kind of evidence out of court.

This article uses the four grades rather than a two-way split, because the difference between bracketed and reconstructed matters, and because the question is always which claim was watched, not which field it belongs to.

Part 2: Cases offered as evolution in action

The cases in one table

Each row grades the claim as it is usually offered.  Where a case has a watched part and a reconstructed part, the table names both.  The last column links the article that works the case in full.

Case What was watched or bracketed What was reconstructed Grade of the usual claim Article
Citrate use in the Long-Term Evolution Experiment The whole event, in a frozen record, and replayed Nothing essential Watched LTEE
Phage lambda uses a new receptor The whole event, in 8 to 15 days Nothing essential Watched Phage Lambda
Directed evolution of enzymes Every round, by design Nothing Watched Directed Evolution
Extended-spectrum beta-lactamases The same substitutions in dated clinical isolates and in a flask Ages of the resistance gene families Watched Antibiotic Resistance
HIV drug resistance Resistance mutations sequenced in treated patients The virus's origin from chimpanzee SIV Watched HIV
Influenza drift and new glycan sites Dated isolates since the 1930s The origins of the 1918 and 2009 viruses Watched Influenza
SARS-CoV-2 variants Over 13 million sequenced genomes The virus's origin Watched SARS-CoV-2
Drug resistance in tuberculosis Resistance arising in treated patients The age of the complex Watched Tuberculosis
Peppered moths turn black and pale again Frequencies from 1848 to today, and predation experiments The date of the black mutation Watched Peppered Moths
Darwin's finch beaks track droughts Every bird measured yearly from 1973 The ages of the beak haplotypes and of the radiation Watched Darwin's Finches
Stickleback armour loss in a lake Yearly samples from 1990 to 2001 The age of the low-armour allele Watched Sticklebacks
Dog breeds Breeding records of 150 to 200 years Ages of many variants Watched Dog Breeds
New plant species by chromosome doubling Found forming in the wild and re-made in gardens How common it was in the past Watched Observed Speciation
Chloroquine resistance in malaria First failures 1957 to 1960 How many times it arose Bracketed Plasmodium falciparum
Lactase persistence in Europe Allele frequencies in dated ancient skeletons The age and origin of each mutation Bracketed Lactase Persistence
EPAS1 in Tibetans Allele frequencies in dated ancient skeletons Arrival from Denisovans Bracketed EPAS1
Italian wall lizards Released 1971, sampled 2004 to 2006 That the change is inherited, which has not been shown Bracketed Italian Wall Lizards
Antifreeze genes in polar fish Nothing Both origins, from genome comparison and a clock Reconstructed Antifreeze Proteins (planned)
De novo and orphan genes Nothing Every case, by comparing genomes Reconstructed Orphan Genes
Cichlid radiations Nothing The radiation, from lake age and clocks Reconstructed Cichlid Radiations
Origin of the finch beak variants Nothing Haplotype splits about a million years old Reconstructed Darwin's Finches
Origin of the black peppered moth Bracketed only by the first record, 1848 Mutation dated about 1819 from recombination Reconstructed Peppered Moths
Origin of the stickleback low-armour allele Nothing Over two million years, from sequence divergence Reconstructed Sticklebacks
London Underground mosquito That the forms do not interbreed The origin, now dated to over 1,000 years Asserted Observed Speciation
Nylonase from a frameshift Bacteria eating nylon waste; a 1995 laboratory selection The frameshift, never tested Asserted Nylonase

Count:  13 watched, 4 bracketed, 6 reconstructed, 2 asserted.  The counts depend on how the cases are split, and a reader could split them differently.  The pattern does not depend on it:  every watched row is small, and every row about the origin of something new is reconstructed.

Watched: the laboratory and the clinic

The Long-Term Evolution Experiment is the cleanest record in biology.  One population gained the ability to use citrate in the presence of oxygen after more than 30,000 generations.  Because every 500th generation was frozen, the team could revive earlier samples and replay the history.2  This is a gain of function by Michael Behe's own published test, and the site concedes it.  See Long-Term Evolution Experiment and Measuring Biological Information.

Justin Meyer's phage lambda populations were likewise watched throughout.  About 0.01% of the phage could infect bacteria lacking the old receptor "by day 8," and "the majority did so by day 15."  The only host resistant to the evolved phage lacked both receptors, "indicating that OmpF was the new receptor."10 10  Four mutations were needed, and each was paid for by selection on the old receptor.  See Phage Lambda and the OmpF Switch.

Directed evolution in the laboratory is watched by construction.  Its limits are about what it can reach, not about whether it happened.  See Directed Evolution.

Antibiotic resistance is watched in two settings.  Miriam Barlow and Barry Hall evolved the TEM beta-lactamase in the laboratory and compared the result with hospitals.  "Nine mutations have arisen independently multiple times in nature," and seven of the nine turned up in their flasks.11  The mutations were watched arising in clinical isolates, and their effects were measured by building the combinations.  Daniel Weinreich's famous five-step landscape was built rather than evolved:  "we first constructed the 32 combinations of these five mutations".12  That is a measurement of the steps, not an observation of a population taking them.  The deep history of resistance genes is a separate question.  Vanessa D'Costa and colleagues recovered working vancomycin-resistance genes from permafrost under an ash layer dated to about 30,000 calendar years.13 13  That is a bracketed observation of an old gene, not a reconstruction.  The ages of whole gene families are reconstructed.  See Antibiotic Resistance.

The fast viruses are watched at a depth no other organism approaches.  By December 2022, 13.4 million SARS-CoV-2 sequences had been deposited in one database.14  Influenza's H3 subtype gained sugar-attachment sites on its surface protein four times after 1968, each dated by stored isolates.15  The origins of these viruses are another matter.  Jeffery Taubenberger's team sequenced the 1918 virus from preserved tissue and could only "propose" that it was "more likely an entirely avian-like virus that adapted to humans".16  The 2009 pandemic virus's ancestors, by a clock, "have been circulating undetected for about a decade."17  The spread is watched and the origin is reconstructed.  The same split holds for HIV, whose resistance mutations are sequenced in treated patients and whose origin from chimpanzee virus is dated by a clock.  See Influenza, SARS-CoV-2 and HIV.

Drug resistance in tuberculosis is watched in treated patients.  The age of the tuberculosis complex, as Part 1 showed, is reconstructed, and two calibrations disagree by more than tenfold.

Watched: animals and plants in the wild

The peppered moth is a watched case with a reconstructed tail.  H. B. D. Kettlewell called industrial melanism "the most striking evolutionary change ever actually witnessed in any organism, animal or plant."18  The rise and fall of the black form were recorded from the first record in Manchester in 1848 to near-extinction of the form today.19  Michael Majerus's final experiment watched birds take the moths:  "approximately 26 per cent were seen being eaten by birds".20  The mutation itself was not watched.  Arjen van't Hof and colleagues found it was a transposable element inserted into the gene cortex.19  They dated the insertion by simulating "One million simulated time trajectories" and comparing recombination around the insert.  The date distribution is "highly skewed (median = 1763, interquartile range = 1681-1806) with a maximum likelihood at 1819".19 19  That the black form descends from "a single mutational event in the recent past" is also inferred, from linkage patterns.21  So the popular sentence "the moth evolved a black form after 1819" joins a watched rise to a reconstructed origin.  The black allele is a new mutation, not an old variant, and Peppered Moths scores it as a modification by copy.

Darwin's finches split the same way.  The Grants recorded that "Significant evolutionary events occurred in G. fortis eight times" and in the cactus finch seven times over 30 years.22  After the 2004 to 2005 drought the change between generations was "0.70 SD," the "strongest evolutionary change seen in the 33 years of the study."23  All of that was watched.  The variants that selection sorted were not.  The two versions of the beak-size region split "about 1 million years ago",24 and the 18 species descend from an ancestor that arrived "∼1.5 million years ago".25  Those are reconstructions from sequence and clocks.  See Darwin's Finches.

Threespine sticklebacks in Loberg Lake, Alaska, were sampled after the lake was restocked.  The 1990 sample "was nearly monomorphic for the complete plate morph," and "By 2001, the frequency of completes had declined to 11%, and lows had increased to 75%."26  That is watched.  The low-armour allele itself belongs to "an ancestral low-plated haplotype that first appeared more than two million years ago",27 a date from sequence divergence.  See Sticklebacks.

Dog breeds were made under written breeding records over 150 to 200 years, so breed formation is watched.  The ages of many of the variants the breeders sorted are reconstructed.  See Dog Breeds.

New plant species formed by doubling their chromosomes have been found forming in the wild and re-made in gardens.  This is the site's first concession on speciation and it is not walked back.  See Observed Speciation.

Bracketed

Chloroquine resistance in the malaria parasite is bracketed.  The drug came into wide use in the early 1950s.  "Anecdotal reports suggest that resistance emerged as early as 1957," and "by 1960, resistance in these areas was confirmed".28  Nobody saw the first resistant parasite arise.  How many times it arose is reconstructed from the DNA around the resistance gene.  John Wootton's team found "at least four CQR founder events."29  Nicholas White put it at "less than ten times in the past fifty years," which gives his famous figure of about one success in 1020 parasites.30  That figure combines a count and a census, both estimated.  See Plasmodium falciparum.

Lactase persistence is bracketed by ancient DNA.  Joachim Burger's team tested early Neolithic Europeans and "did not observe the allele most commonly associated with LP in Europeans".31  Iain Mathieson and Jonathan Terhorst traced its rise in dated British genomes and found selection "strongest (s ≈ 0.08) from 150 to 100 generations before the present".32  The frequency change is bracketed by dated skeletons.  The mutation's origin is reconstructed.  See Lactase Persistence.

The Tibetan EPAS1 variant is bracketed the same way.  Choongwon Jeong's team found reads for the derived variants "in two of the three later Samdzong individuals" but not in the earlier ones.33  Its arrival from Denisovans is reconstructed from its pattern of shared differences:  the haplotype structure "can only be convincingly explained by introgression".34  See EPAS1 and the Tibetan Highlands.

The Italian wall lizards of Pod Mrčaru are bracketed, and the bracket is the whole of the evidence.  Ten lizards were released in 1971.  Biologists returned in 2004 to 2006 and found bigger heads and gut valves.  Richard Dawkins called it "evolution before our very eyes."35  The discoverers were more careful.  They wrote that "further studies investigating the potential role of phenotypic plasticity and/or maternal effects in the divergence between populations are needed."36  Later work found no valves in "any of the twenty specimens that had been fed arthropods",37 and head and bite differences "did not persist in a common garden environment."38  A change was bracketed.  That it was evolution, in the sense of an inherited change, has not been shown.  Answers in Genesis predicted in 2009 that the valve "may dwindle" on an insect diet, and it did.39  See Italian Wall Lizards.

Reconstructed

The antifreeze genes of polar fish are the series' third concession:  the Arctic cod gene arose from DNA that was not a gene.  Nobody saw it happen.  Xuan Zhuang and colleagues "constructed genomic DNA libraries for AFGP-bearing and AFGP-lacking species across the gadid phylogeny and performed fine-scale comparative analyses".40  Their aim was to "reconstruct the gadid AFGP evolutionary path."  The Antarctic gene came from a trypsinogen gene, and its date comes from a clock:  "The small sequence divergence (4–7%) between notothenioid AFGP and trypsinogen genes indicates that the transformation of the proteinase gene into the novel ice-binding protein gene occurred quite recently, about 5–14 million years ago".41  A later study placed the origin on a fossil-calibrated family tree and found that "the origin of AFGP occurred between 42 and 22 Ma",42 so the date has moved;  both figures are inferences.  The site concedes the Arctic case as a real gain, and the concession rests on reconstruction alone.  The evidence is strong of its kind, because the founding non-coding sequence can still be seen in relatives that lack the gene.  It is still comparative evidence.

De novo genes are reconstructed in every case.  Dong-Dong Wu and colleagues reported 60 human genes that arose from non-coding DNA and a rate of about 9.83 to 11.8 genes per million years.  The rate depends on the split from chimpanzees:  "Since the chimpanzees and humans shared a common ancestor" 5 to 6 million years ago, the count divided by that span gives the rate.43  Whether a gene without relatives is new, or old and diverged past detection, is the question Orphan Genes works through.  Caroline Weisman and colleagues found that many lineage-specific genes can be explained by failure to detect their relatives.44  Protein-rarity arguments do not apply to these short sequences, and the site does not aim them there.

The cichlid radiations of the African lakes were not watched.  Their speed is inferred from the age of the lakes and from clocks.  See Cichlid Radiations.

Asserted as observed

The London Underground mosquito was the most widely taught case of evolution in a human-built habitat.  The original paper hedged.  Katharine Byrne and Richard Nichols found that "no offspring were recorded from crosses between Underground and surface populations," and concluded only that "the reproductive isolation could have accumulated in the few hundred generations since then."45 45  In retelling it lost its "could have."  Carl Wieland of what is now Creation Ministries International wrote that the mosquitoes entered the Tube tunnels "About 100 years ago" and that the evidence was "thus suggesting that they have become a new species".46  In 2025 Yuki Haba and colleagues sequenced about 350 modern and historical samples.  Their work "counters the popular hypothesis that molestus originated belowground in London <200 years ago".47  They noted that "the hypothesis of in situ urban evolution has never been rigorously tested" and that their work "debunks one of the most widely cited examples of rapid urban adaptation".47 47  Their own replacement date is a reconstruction too.  It needs a mutation rate and generation time "neither of which has been directly measured for Cx. pipiens in nature", and it spans "1300 and 12,500 years ago".47  What stands as observed is that the two forms do not interbreed.  See Observed Speciation.

Nylonase is the other textbook case.  In 1984 Susumu Ohno suggested that the enzyme arose by a frameshift from an older coding sequence.  He wrote "I assume" and "I propose", and his ancestral sequence was built "if T marked by an arrow in the 3rd row of Fig. 1a is deleted".48 48  A year later William Thwaites reported in the National Center for Science Education's journal:  "Now it has happened!  Not one, but two, new proteins have been discovered."49  The frameshift was never tested.  Later work found the enzyme belongs to an old esterase family, and that two of 47 amino acid differences account for its nylon activity.50 51  The bacteria that eat nylon waste are real, and a 1995 laboratory selection watched the ability appear, though its genetics were never published.  The frameshift is an assertion.  See Nylonase.

The TalkOrigins Archive's "Observed Instances of Speciation" is the list most often cited online.  It mixes grades.  Its plant polyploids are watched.  Its entry for the apple maggot fly adds its own warning that "some of the leading researchers on this question are urging caution in interpreting it."52  A list is only as good as its worst row, and Observed Speciation grades each row.

Part 3: The same test on the design and creationist side

Claim What was observed What was inferred or modelled Grade Article
Axe: working enzyme folds "may be as low as 1 in 1077" Mutational tolerance of one deliberately weakened enzyme The extrapolation to all sequences and all folds Reconstructed How Rare Is a Working Protein?
"Only 1 in 1077 sequences are functional" Nothing beyond the above Stated as a finding Asserted same
Dembski: the flagellum's probability is about 10-234 Nothing A random draw of proteins, a process no biologist proposes Reconstructed (modelled) The Design Inference
Behe: no stepwise route to the clotting core or the flagellum has been shown True: none has been rebuilt with each step's benefit measured Nothing Watched (an observed absence) Irreducible Complexity
Behe: such systems could not arise stepwise The absence above That no route exists Reconstructed Blood Clotting, Bacterial Flagellum
"There are no beneficial mutations" Contradicted by watched cases Nothing Asserted Measuring Biological Information
"Mutation and natural selection only lead to a loss of functional systems" Contradicted by watched gains Nothing Asserted same
No antibiotic-destroying enzyme has been watched arising from DNA that was not a gene True as of now Nothing Watched (an observed absence) Antibiotic Resistance
The lizard valves would dwindle on an insect diet Confirmed in captive lizards Nothing Watched (a tested prediction) Italian Wall Lizards
Rapid post-Flood speciation within created kinds Fast speciation from existing variation, in cichlids and finches The kinds, the founding pairs and the 4,400-year timetable Reconstructed Creation Model of Rapid Speciation
Genetic entropy: human fitness declining for 6,000 years Mutation rates per generation The decline and its history, from models Reconstructed (modelled) Genetic Entropy
Parasites and pathogens were harmless before the Fall Some pathogens harm only when displaced; some symbioses break down An original harmless state Asserted Predation, Parasites, and Pathogens
Chloroquine resistance needed two mutations together, at a cost of about 1020 parasites Later laboratory work on the resistance gene The per-parasite rate Watched in part Plasmodium falciparum

Count:  4 rest on observation (three observed absences or tested predictions, plus the chloroquine reading), none bracketed, 5 reconstructed or modelled, 4 asserted.

Axe's 1 in 1077

Douglas Axe's 2004 paper measured how many random changes a weakened bacterial enzyme could tolerate and still work.  His method was "to obtain an extensively degraded reference sequence that just passes a low selection threshold" and then randomise small parts of it.53  From that measurement and two further estimates he concluded that the share of sequences "performing a specific function by any domain-sized fold may be as low as 1 in 1077".53  That is a reconstruction.  It extrapolates from one enzyme, one fold and one weakened variant to all of sequence space, and Axe's own wording carries the hedge.

In 2023 Brian Miller and Casey Luskin of the Discovery Institute restated it.  First:  Axe "found that the likelihood of a chance sequence of 153 amino acids generating the stable, functional fold needed for the larger domain in that enzyme was as low as 1 in 1077."  Then, quoting their own paper:  "Axe (2000, 2004)’s work on the larger beta-lactamase protein domain indicates that only 1 in 1077 sequences are functional".54  "May be as low as" became "found" and then "only."  "A specific function by any domain-sized fold" became "sequences are functional."  The measurement did not change.  The grade did.  How Rare Is a Working Protein? works the numbers.

Dembski's flagellum

William Dembski's 2002 book No Free Lunch priced the bacterial flagellum.  Its one worked number is the chance that "Randomly picking 250 proteins" gives the right ones, which "has order of magnitude 10-234".55  No biologist proposes that the flagellum was assembled by a random draw of proteins, so the number models a process nobody holds.  Dembski himself set it aside.  In 2005 he wrote that the needed probability, under a chance hypothesis that includes "Darwinian and other material mechanisms," "has yet to be done."56  In 2008 he wrote:  "The example in NFL ch. 5 doesn't fit the bill."57  A reader who cites the number as the flagellum's probability is asserting something its author withdrew.  See The Design Inference.

Behe's irreducible complexity

Michael Behe's claim has two parts, and they have different grades.  The first is an observation about the record.  Nobody has rebuilt a stepwise route to the clotting core or the flagellum with the benefit of each step measured.  Behe wrote in 1996 that "no one on earth has the vaguest idea how the coagulation cascade came to be."58  Russell Doolittle, who built the leading evolutionary account, wrote in 2009 that the lack of any living rudimentary system leaves "us to speculate on how the system might have gotten started."59  So the evolutionary history of clotting is a reconstruction, and its author says so.

The second part is that such systems could not arise stepwise.  That is an inference, and Behe hedged it himself:  "one can not definitively rule out the possibility of an indirect, circuitous route. As the complexity of an interacting system increases, though, the likelihood of such an indirect route drops precipitously."58  "Drops precipitously" is a probability claim, and nobody has computed it for any named machine.  Both sides here are reconstructing.  See Irreducible Complexity, Blood Clotting and Bacterial Flagellum.

What the ministries say about mutations

Creation Ministries International lists "There are no beneficial mutations" among arguments creationists should not use:  "This is not true, since some changes do confer an advantage in some situations. Rather, we should say, ‘We have yet to find a mutation that increases genetic information in ways needed for microbes-to-man evolution to be feasible."60  The first half is an observation, and it is right.  The second half is a claim about extrapolation.  It is where the real argument lies, and it is not an observation that settles anything.

Other pages go further.  Georgia Purdom wrote for Answers in Genesis in 2007 that "Mutation and natural selection, thought to be the driving forces of evolution , only lead to a loss of functional systems."61  That is an assertion, and watched cases contradict it:  citrate use in the Long-Term Evolution Experiment, the new receptor in phage lambda, the extended-spectrum beta-lactamases and the new sugar sites on influenza.  None of these is a loss of a functional system.

The narrow claim survives.  Don Batten of Creation Ministries International wrote in 2017 that "in the instances known, the machinery to manufacture the antibiotic-destroying enzyme has never been observed to arise by mutation."62  That is a statement about the watched record, and Antibiotic Resistance found it true.  It is the right form for a creationist claim:  specific, checkable and about what was observed.

Post-Flood speciation and created kinds

The creation model holds that animals diversified fast after the Flood from the pairs on the Ark.  Fast speciation from existing variation is watched, in the Big Bird finch lineage and in plants, so the mechanism has observational support.1  The kinds, the founding pairs and the timetable are reconstructed.  Todd Wood said so plainly in 2003:  "Since we see no ‘speciation’ like this today or in historical times, we can infer that the mechanism of intrabaraminic diversification is qualitatively different from modern speciation."63  His model requires "the origin of 150 species of horse" within about 400 years of the Flood.63  That is a reconstruction from Scripture and from the living species, of exactly the kind creationists call historical science when evolutionists do it.  See The Creation Model of Rapid Speciation.

Genetic entropy

John Sanford's Genetic Entropy shows a curve of human fitness falling 1% per generation for 300 generations, 6,000 years.  The caption says that "Dr. Crow (1997) indicates that he believes the fitness of the human race is presently degenerating at 1-2% per generation," and that the curve is "shown for a hypothetical human population".64  James Crow's sentence was conditional and modern:  "If we are like Drosophila, the decrease in viability from mutation accumulation is some 1 or 2% per generation. This is more than compensated for by much more rapid environmental improvements".65  Crow was describing the effect of relaxed selection in recent centuries.  Running his figure back 6,000 years reconstructs a history from a model, and no measurement of human fitness over that span exists.  The mutation counts per generation are measured.  The decline is modelled.  See Genetic Entropy.

Harmless before the Fall

Young-earth writers hold that disease-causing microbes and parasites were once harmless.  Joseph Francis set out two routes:  "(1) post-Fall genetic alteration of the original good microbe and/or (2) post-Fall displacement or movement of the microbe from the site where it performed its beneficial function."66  The displacement half rests on observed ecology.  The cholera bacterium's virulence factors also help it grow on crustacean shells in the sea.  The original harmless state has no dated observation and no reconstruction from sequence that begins with a harmless organism.  It is asserted from Genesis.  That can be a legitimate theological claim.  It is not an observation, and Predation, Parasites, and Pathogens finds the evidence mixed.

Part 4: The pattern

The counts.  On the evolutionary side, 13 cases are watched, 4 bracketed, 6 reconstructed and 2 asserted as observed.  On the design and creationist side, 4 claims rest on observation, none are bracketed, 5 are reconstructed or modelled and 4 are asserted.

What the watched cases share.  Every watched evolutionary case is small.  By the scoring in Benchmarking Evolution, they are mostly modifications and losses, with a few gains of one to five mutations:  citrate use, the lambda receptor, the beta-lactamases, influenza's sugar sites and chloroquine transport.  They happen within one species or one population.  They take days to two centuries.  Several are shifts in how common an existing variant is, as with the finches, the sticklebacks and the dogs.

What the reconstructed cases share.  They are the large gains:  a gene from non-coding DNA, a new antifreeze from a digestive enzyme, the origin of the variants that selection later sorted, and radiations of hundreds of species.  Each rests on a tree, a clock or both.

What follows.  The large gains being reconstructed does not make them false.  It means the evidence for them is comparative, and it should be judged by the methods for comparative evidence:  whether independent trees agree, whether clocks calibrated independently agree, and whether a founding sequence can still be seen in relatives, as it can for the Arctic cod antifreeze.  Phylogenetic Congruence, Orphan Genes and the fossil articles do that work.  What the pattern does support is the series' argument about extrapolation:  the watched record shows what small searches buy, and the claim that the same process built the large gains is an inference from that record, not an observation in it.

The same holds on the other side.  A design argument resting on inference is not therefore false.  Behe's "likelihood of such an indirect route drops precipitously" may be right.  But it is a reconstruction, like Doolittle's scenario, and it needs a calculation neither side has done.

The hedge that goes missing.  In each case of inference sold as observation, the originator hedged and a later writer dropped the hedge:

Originator's words Later words
Byrne and Nichols: isolation "could have accumulated" Wieland: "they have become a new species"
Ohno: "I propose" Thwaites: "Now it has happened!"
Axe: "may be as low as 1 in 1077" Miller and Luskin: "only 1 in 1077 sequences are functional"
Crow: "If we are like Drosophila" Sanford: 6,000 years of human decline
Herrel: plasticity studies "are needed" Dawkins: "evolution before our very eyes"

Objections and responses

"Inference is how all science works.  Nobody watches an electron either."  True, and this article does not demote inference.  Its point is narrower.  A change whose before and after were recorded does not depend on a tree or a clock, and a reconstructed change does.  The section of this site on observed evolution admits only the first kind, so that its numbers do not carry assumptions from the second.

"Common descent is supported by consilience, so its reconstructions are as good as observations."  Some are very strong.  The founding sequence of the Arctic cod antifreeze gene is still visible in cod relatives, which is close to a smoking gun in Cleland's sense.  Others rest on one calibration, as the tuberculosis dates do, and those can be wrong by a factor of ten.  Each reconstruction has to be judged on its own checks.  Consilience is a reason to trust a well-tested one, not a reason to call it watched.

"You weren't there" is a creationist evasion.  Used alone it is, and this article does not use it that way.  The same test is applied to the creation model, and the kinds, the post-Flood timetable and the harmless pre-Fall microbes fail it.  Geisler and Anderson said as much in 1987.

"The peppered moth is a shift between existing alleles, so it is not evolution."  The black form is a new mutation, a transposable element in cortex, dated by reconstruction to around 1819 and first recorded in 1848.  Its rise and fall were watched.  It is evolution by any definition, and a small one.

"Watched cases are all microbes, so the pattern is biased."  The finches, moths, sticklebacks, dogs and polyploid plants are watched animals and plants, and they show the same pattern:  selection on variants that were already present or on one new mutation.

"The design side is held to a lower standard, since four of its claims are counted as observations."  Three of the four are observations of absence or a tested prediction, which are the weakest kind of observation.  The table grades the design side's two most cited numbers, Axe's and Dembski's, as reconstruction and assertion.

"If the large gains are only reconstructed, the case for evolution collapses."  No.  Reconstruction is evidence.  The argument this site makes is about extrapolation and the rate of gain, set out in Benchmarking Evolution.  It does not follow from this audit that the reconstructions are wrong.

Issues and unanswered questions

  • Several gains are watched. Citrate use, the lambda receptor, the beta-lactamases and the influenza sugar sites are gains of function watched in real time. Any young-earth argument that says gains are never observed is wrong, and the site concedes them.
  • The strongest reconstructions are hard to answer. A founding non-coding sequence still visible in relatives, as in the Arctic cod, is the kind of trace Cleland means. No young-earth author has published an account of why the cod's antifreeze gene sits where a non-coding relative predicts.
  • The young-earth model's own core is reconstructed. The created kinds and the post-Flood timetable are inferred from Scripture and from living species. No observation brackets them. The creation model cannot claim the evidential advantage of "observational science" for its own history.
  • Some bracketed cases may yet become watched or fall away. The lizards need a test of inheritance of the valves. The date of the black peppered moth could be tested by genotyping moths collected before 1848, if any survive in museums.
  • Grading is a judgement. A reader could split the cases differently or move a row between grades. The counts are a summary, not a measurement.
  • Cleland's paper was read in abstract only. Her definition of a smoking gun is quoted through Forber and Griffith. Her full argument, including her examples, was not checked against her text.

Verdict

What was watched is real, small and within species:  one to a few mutations, or a change in how common an existing variant is, over days to two centuries.  What is large, the origin of new genes and new structures and the radiations, was reconstructed from comparisons, and some of it has been presented as observed when it was not.  That supports this site's argument that the watched record does not by itself show the large gains, and it does not show that the reconstructions are wrong.  The design and creationist side has made the same error:  a hedged extrapolation sold as a measurement, a withdrawn calculation still cited, and a model of human decline sold as history.  Its own account of origins, the kinds and the post-Flood timetable, is a reconstruction under its own rules.  The honest position on both sides is to say which claims were watched and to argue about the others on the strength of their methods.

Sources

  1. 1.

    Lamichhaney, Sangeet et al.  “Rapid hybrid speciation in Darwin’s finches.”  Science 359(6372):224–228.  2018.  Only the abstract was read.  The body is behind a login.  Copyright.

    1. 1a.

      Page 224, abstract:  “Homoploid hybrid speciation in animals has been inferred frequently from patterns of variation, but few examples have withstood critical scrutiny. Here we report a directly documented example, from its origin to reproductive isolation.” ↩ ↩ ↩

  2. 2.

    Blount, Zachary D. et al.  “Historical contingency and the evolution of a key innovation in an experimental population of Escherichia coli.”  Proceedings of the National Academy of Sciences 105:7899–7906.  2008.  Read from the publisher PDF in the library.  PDF page 2 carries the printed page number 7900.

    1. 2a.

      Page 7899:  “samples have been frozen every 500 generations, providing a rich “fossil record””. ↩ ↩

  3. 3.

    Cleland, Carol E.  “Methodological and Epistemic Differences between Historical Science and Experimental Science.”  Philosophy of Science 69(3):474–496.  2002.  Cleland is a philosopher of science at the University of Colorado.  Read in abstract only, from the publisher's page.  The full text is closed:  no repository copy, and PhilPapers refused the download.  Copyright.

    1. 3a.

      Section Abstract:  “Historical science (which includes work in geology, biology, and astronomy, as well as paleontology and archaeology) is sometimes held to be inferior on the grounds that its hypothesis cannot be tested by controlled laboratory experiments.”  “It rejects the claim that historical research is epistemically inferior.”  “I show that these different patterns of reasoning are grounded in an objective and remarkably pervasive time asymmetry of nature.” ↩

  4. 4.

    Forber, Patrick, and Eric Griffith.  “Historical Reconstruction: Gaining Epistemic Access to the Deep Past.”  Philosophy and Theory in Biology 3:e203.  2011.  Read from the Internet Archive copy of the publisher's full text, because the publisher's site blocks scripted downloads.  CC BY-NC-ND 3.0.  The HTML has no page numbers.  Mirror:  Archive.org

    1. 4a.

      Section Notes:  “A smoking gun is a trace(s) that unambiguously discriminates one hypothesis from among a set of currently available hypotheses as providing ‘the best explanation’ of the traces thus far observed.”  In the endnote attached to section 4's remark that Cleland takes several traces together as a smoking gun.  Forber and Griffith are quoting Cleland 2002, pp. 480–481, whose full text was not read. ↩

    2. 4b.

      Section 2. The Problem of Access:  “with no way to test any claim against the actual event in real time, how can epistemic support accrue to historical claims such that they become more than mere just-so stories?” ↩

    3. 4c.

      Section Abstract:  “We argue that strong epistemic support for historical claims consists in the consilience of multiple independent lines of evidence”.  “the general conditions for this are strict”. ↩

  5. 5.

    Comas, Iñaki et al.  “Out-of-Africa migration and Neolithic coexpansion of Mycobacterium tuberculosis with modern humans.”  Nature Genetics 45(10):1176–1182.  Oct 2013.  PMC author manuscript read through bin/pmcget.py. No printed folios, so locators are section names. Reference superscripts are removed from the quotations.  Mirror:  PMC

    1. 5a.

      Section Results:  “Because of these limitations, we used an alternative approach to date our MTBC phylogeny. Specifically, we used as initial calibration points several key dates in human evolution.”  Three models were tested, tying the basal split to anatomically modern humans at 185 thousand years, to the mitochondrial L3 group at 70 thousand years, or to the start of the Neolithic at 10 thousand years. The 70 thousand year model is the one the paper prefers. ↩

    2. 5b.

      Section Discussion:  “The accumulation of more than 30 thousand SNPs by human MTBC strains over the proposed timeframe of 70 thousand years corresponds to a long-term genome-wide substitution rate of 2.58 × 10−9 substitutions per site per year (95% HPD 1.66 × 10−9 to 2.89 × 10−9, Table 1).” ↩

    3. 5c.

      Section Results:  “Moreover, when performing a tip-to-date analysis of those strains (N = 49), we found that in contrast to several other bacterial species, no significant correlation between isolation time and phylogenetic divergence exists in MTBC (correlation coefficient = 0.047).” ↩

  6. 6.

    Bos, Kirsten I. et al.  “Pre-Columbian mycobacterial genomes reveal seals as a source of New World human tuberculosis.”  Nature 514(7523):494–497.  Aug 20, 2014.  PMC author manuscript read through bin/pmcget.py. No printed folios, so locators are section names. The three ancient genomes are from radiocarbon-dated Peruvian skeletons and belong to the seal-adapted lineage, which Sabin 2020 names as the weakness in this calibration.  Mirror:  PMC

    1. 6a.

      Section Abstract:  “Two independent dating approaches suggest a most recent common ancestor for the M. tuberculosis complex less than 6,000 years ago, which supports a Holocene dispersal of the disease.” ↩

  7. 7.

    Geisler, Norman L., and J. Kerby Anderson.  Origin Science: A Proposal for the Creation-Evolution Controversy.  Baker Book House.  1987.  Geisler was a Christian philosopher and apologist;  Anderson a Probe Ministries writer with a background in biology.  Read on archive.org item originsciencepro0000geis, which cannot be borrowed, by paragraph search.  Not stored.  Printed page = leaf minus 2, from the item's scan data, checked at leaves 15, 27, 100 and 150.  Copyright.

    1. 7a.

      Page 14:  “It is the proposal of this book that a science which deals with origin events does not fall within the category of empirical science, which deals with observed regularities in the present. Rather, it is more like a forensic science, which concentrates on unobserved singularities in the past.”  The scan hyphenates "sin- gularities" across a line. ↩

    2. 7b.

      Page 25:  “In the strict scientific sense, it is true that special creation is not falsifiable. But neither is macroevolution.”  “Microevolution is an empirical science, since it can be observed in the present”.  The second quotation is the page's footnote. ↩

  8. 8.

    Lacey, Troy.  “Deceitful or Distinguishable Terms—Historical and Observational Science.”  Answers in Genesis, Feedback.  Jun 10, 2011.  Retrieved Sep 29, 2026.  A reader's letter and Lacey's answer on behalf of Answers in Genesis.  Copyright.

    1. 8a.

      Section Reader's letter:  “Or is “historical” simply anything which contradicts and “observational” anything that supports you?” ↩

    2. 8b.

      Section Answer:  “we have stated that neither creationism nor cosmic evolution nor Darwinian biological evolution is observational science, and they are not observable, testable, repeatable, falsifiable events.” ↩

  9. 9.

    Mitchell, Elizabeth.  “Four-Legged Snake Fails as a Transitional Form and Perhaps Even as a Snake.”  Answers in Genesis, News to Know.  Nov 22, 2016.  Copyright.

    1. 9a.

      Section Unobservable, Untestable, Unrepeatable:  ““Origins science” is not observational science.”  “Evolutionary transitions from one kind of animal to another have never been observed.” ↩

  10. 10.

    Meyer, Justin R. et al.  “Repeatability and Contingency in the Evolution of a Key Innovation in Phage Lambda.”  Science 335(6067):428–432.  Jan 27, 2012.  Read as the PubMed Central author manuscript (NIHMS362453), both the HTML and the 14-page PDF, which carry identical text.  Every page below is an author-manuscript page read from the "Meyer et al. Page N" running head, not a Science page:  the published pagination could not be obtained, since the Science PDF is paywalled and Unpaywall and Semantic Scholar list only the PubMed Central copy.  Exponents are restored from the sense of the text, because the PDF and HTML both flatten superscripts.  Mirror:  PMC

    1. 10a.

      Section the paragraph on the six replicate flasks cultured for 28 days:  “Such spots were observed in only one population, but this ability evolved quickly, such that ~0.01% of the phage could infect the lamB-negative mutants by day 8, including the isolate designated EvoC, and the majority did so by day 15.” ↩

    2. 10b.

      Section the paragraph on the seven knockout strains:  “The only host that was resistant to the EvoC isolate was the ompF malT double mutant that lacked expression of both OmpF and LamB (Fig. 1, table S2), indicating that OmpF was the new receptor.” ↩

  11. 11.

    Barlow, Miriam, and Barry G. Hall.  “Predicting Evolutionary Potential: In Vitro Evolution Accurately Reproduces Natural Evolution of the TEM β-Lactamase.”  Genetics 160(3):823–832.  Mar 2002.  Read from the PubMed Central PDF; PDF page N is printed folio 822 plus N, checked at the title page and at the substitutions paragraph.  Mirror:  PMC

    1. 11a.

      Page 829:  “Nine mutations have arisen independently multiple times in nature: L12F, E104K, R164S, R164H, M182T, A237T, G238S, E240K, and T265M.  Seven of those, E104K, R164S, R164H, M182T, G238S, E240K, and T265M, were recovered during our in vitro evolution experiments”. ↩

  12. 12.

    Weinreich, Daniel M. et al.  “Darwinian Evolution Can Follow Only Very Few Mutational Paths to Fitter Proteins.”  Science 312(5770):111–114.  Apr 7, 2006.  Paywalled at the publisher and reported closed by Unpaywall, OpenAlex, Europe PMC and Semantic Scholar's public API. The full text was obtained from a Wayback capture of a university course mirror of the publisher PDF, with printed folios 111 to 114 in the running feet. The library previously held the abstract only. The PDF interleaves with an unrelated economics Report, so each page was extracted separately, and Table 1 was read off a 110 dpi page image because pdftotext destroys its exponents.  Mirror:  Archive.org

    1. 12a.

      Page 111:  “To characterize the effect on drug resistance of each mutation on all allelic backgrounds, we first constructed the 32 combinations of these five mutations”.  The alleles were built, not evolved. No population size and no waiting time were measured. ↩

  13. 13.

    D'Costa, Vanessa M. et al.  “Antibiotic resistance is ancient.”  Nature 477(7365):457–461.  Aug 31, 2011.  Paywalled at the publisher and absent from every open aggregator checked. Read from a Wayback capture of the publisher's full-text HTML made from a subscribing library. The HTML has no print folios, so locators here are the abstract and numbered paragraphs of the main text. The Supplementary Information was not obtained.  Mirror:  Archive.org

    1. 13a.

      Paragraph 2:  “The tephra has been dated at several sites in the area to about 25,300 radiocarbon (14C) years BP, or about 30,000 calendar years”.  Main text, second numbered paragraph. The sediment itself was not radiocarbon dated; the age is stratigraphic, from an overlying ash layer. ↩

    2. 13b.

      Paragraph 14:  “These results clearly show that the vanHAX genes identified in the ancient samples encode enzymes capable of genuine antibiotic resistance.” ↩

  14. 14.

    Ip, Jonathan Daniel et al.  “Global prevalence of SARS-CoV-2 3CL protease mutations associated with nirmatrelvir or ensitrelvir resistance.”  eBioMedicine 91:104559.  May 2023.  First published online 14 April 2023.  University of Hong Kong.  CC BY-NC-ND.  Read from the PMC HTML;  locators are section names.  Mirror:  PMC

    1. 14a.

      Section Summary, Methods:  “determined their prevalence among 13.4 million sequences deposited in GISAID as of December 14, 2022, about 1 year after the approval of nirmatrelvir-ritonavir.”  The one published GISAID SARS-CoV-2 count this article could verify.  The exact denominator printed in the Findings is 13,446,588. ↩

  15. 15.

    Altman, Meghan O. et al.  “Human influenza A virus hemagglutinin glycan evolution follows a temporal pattern to a glycan limit.”  mBio 10(2):e00204-19.  Apr 2, 2019.  Open access.  Yewdell is at the National Institute of Allergy and Infectious Diseases;  parts of the work are US Government work.  Read from the Europe PMC full-text XML.  eLife-style article with no printed pages, so locators are section and figure-legend names.  Glycan counts by gel migration are measurements;  per-year frequencies are NetNGlyc predictions and the paper labels them so.  Mirror:  PMC

    1. 15a.

      Section Results, H3 HA glycan evolution:  “The original (1968) pandemic H3 strains had 2 head glycans.  Four times since 1968, glycans were added to H3 HA at regular intervals of 5 to 7 years (N126 by 1974, N246/N122 by 1980, N133-plus-N122 replacement of N276 by 1998, and N144 by 2004)”. ↩

  16. 16.

    Taubenberger, Jeffery K. et al.  “Characterization of the 1918 influenza virus polymerase genes.”  Nature 437(7060):889–893.  Oct 6, 2005.  Copyright Nature;  excerpt only.  The publisher serves the PDF free at https://www.nature.com/articles/nature04230.pdf with a browser user-agent.  Read with pdftotext -layout, which interleaves the two columns, so every quotation was reassembled across columns and re-checked.  Page numbers are the PDF's own running heads.

    1. 16a.

      Page 889:  “we present sequence and phylogenetic analyses of the complete genome of the 1918 influenza virus, and propose that the 1918 virus was not a reassortant virus (like those of the 1957 and 1968 pandemics), but more likely an entirely avian-like virus that adapted to humans”. ↩

  17. 17.

    Smith, Gavin J. D. et al.  “Origins and evolutionary genomics of the 2009 swine-origin H1N1 influenza A epidemic.”  Nature 459(7250):1122–1125.  Jun 25, 2009.  Copyright Nature;  excerpt only.  The publisher serves the PDF free at https://www.nature.com/articles/nature08182.pdf with a browser user-agent.  Read with pdftotext -layout, which interleaves the two columns, so every quotation was reassembled across columns and re-checked.  Page numbers are the printed folios.

    1. 17a.

      Page 1123:  “We found that the common ancestor of the S-OIV outbreak and the closest related swine viruses existed between 9.2 and 17.2 years ago, depending on the genomic segment, hence the ancestors of the epidemic have been circulating undetected for about a decade.” ↩

  18. 18.

    Kettlewell, H. B. D.  “Selection experiments on industrial melanism in the Lepidoptera.”  Heredity 9(3):323–342.  1955.  Scanned PDF from nature.com, read by plain curl.  PDF page n is printed page 322 + n, anchored on the running heads.  The text layer's numerals are unreliable, so every count was checked by arithmetic.

    1. 18a.

      Page 323:  “The industrial melanism of the Lepidoptera is the most striking evolutionary change ever actually witnessed in any organism, animal or plant.” ↩

  19. 19.

    van't Hof, Arjen E. et al.  “The industrial melanism mutation in British peppered moths is a transposable element.”  Nature 534(7605):102–105.  2016.  Read from the accepted manuscript with figures in the University of Liverpool repository, 20 pages.  Page numbers below are manuscript pages, not Nature pages.  Mirror:  Livrepository.liverpool.ac.uk

    1. 19a.

      Section author manuscript, page 3:  “1848 Manchester is generally regarded as the first reported sighting of the carbonaria form, although the wording of the record implies that it was rare but not completely unknown at this time.”  A superscript reference number after "form" is omitted. ↩

    2. 19b.

      Section author manuscript, page 3:  “The probability density for the date of the carb-TE mutation event (Fig. 2d) is highly skewed (median = 1763, interquartile range = 1681-1806) with a maximum likelihood at 1819, a date highly consistent with a detectable frequency being achieved in the mid-1840s.” ↩

    3. 19c.

      Section author manuscript, page 3:  “One million simulated time trajectories of the carbonaria phenotype were randomly drawn according to their fit to historical frequency data (Extended Data Fig. 4). Based on these trajectories, recombination patterns were simulated using an empirical estimate of recombination rate and compared to the observed recombination pattern of the carbonaria haplotypes.”  Section author manuscript, page 5:  The Methods give the recombination rate as 2.9 cM/Mb, estimated from 350 offspring in 8 crosses;  page 6 repeats the rate for the ageing analysis. ↩

    4. 19d.

      Section abstract:  “Here we show that the mutation event giving rise to industrial melanism in Britain was the insertion of a large, tandemly repeated, transposable element (TE) into the first intron of the gene cortex.”  Section main text:  Read from the 20-page author manuscript, which has no printed page numbers.  Its second page reports the 87 candidate differences and their elimination down to the one insert. ↩

  20. 20.

    Cook, L. M. et al.  “Selective bird predation on the peppered moth: the last experiment of Michael Majerus.”  Biology Letters 8(4):609–612.  2012.  Library PDF.  PDF page n is printed page 608 + n.  Data deposited at Dryad, doi 10.5061/dryad.962262h9.  Mirror:  PMC

    1. 20a.

      Page 610:  “Of those that disappeared, approximately 26 per cent were seen being eaten by birds via binocular observations”.  Page 611, table 2:  The caption of Table 2 names the nine bird species. ↩

  21. 21.

    Cook, L. M., and I. J. Saccheri.  “The peppered moth and industrial melanism: evolution of a natural selection case study.”  Heredity 110(3):207–212.  2013.  Review by two of the moth's main field and genetic workers.  Library PDF;  PDF page n is printed page 206 + n.

    1. 21a.

      Page 210:  “Patterns of linkage disequilibrium around the melanism locus indicate that, throughout the United Kingdom, the melanic morph carbonaria is descended from a single mutational event in the recent past (Van’t Hof et al., 2011), thus settling the question whether industrial melanism arose several times in Britain and showing that the initial extension of range was a result of migration.” ↩

  22. 22.

    Grant, Peter R., and B. Rosemary Grant.  “Unpredictable evolution in a 30-year study of Darwin's finches.”  Science 296(5568):707–711.  2002.  Library PDF.  PDF page n is printed page 706 + n, read from the folio on each page.

    1. 22a.

      Page 708:  “Significant evolutionary events occurred in G. fortis eight times (body size, four; beak size, three; beak shape, one) and in G. scandens seven times (body size, two; beak size, five).” ↩

  23. 23.

    Grant, Peter R., and B. Rosemary Grant.  “Evolution of character displacement in Darwin's finches.”  Science 313(5784):224–226.  2006.  Read in the science.org PDF, free with registration, and cross-checked against the archive.org microfilm sim_science_2006-07-14_313_5784 (leaf 97 is page 224 by the item's scan data).

    1. 23a.

      Page 226:  “The difference between generations is 0.70 SD, which is exceptionally large”.  “This is the strongest evolutionary change seen in the 33 years of the study.” ↩

  24. 24.

    Lamichhaney, Sangeet et al.  “A beak size locus in Darwin's finches facilitated character displacement during a drought.”  Science 352(6284):470–474.  2016.  science.org PDF, free with registration.  Folios 470, 472 and 473 read on PDF pages 1, 3 and 4.

    1. 24a.

      Page 473:  “The split between the two haplotypes occurred before the divergence of warbler and nonwarbler finches at the base of the phylogeny (Fig. 1D), about 1 million years ago”.  The region is described on the same page as "this ~525-kb region". ↩

  25. 25.

    Han, Fan et al.  “Gene flow, ancient polymorphism, and ecological adaptation shape the genomic landscape of divergence among Darwin's finches.”  Genome Research 27(6):1004–1015.  2017.  Read in the PubMed Central HTML;  locators are sections.  Mirror:  PMC

    1. 25a.

      Section Introduction:  “18 species of Darwin's finches evolved from a common ancestor that colonized the Galápagos ∼1.5 million years ago”. ↩

  26. 26.

    Bell, Michael A. et al.  “Twelve Years of Contemporary Armor Evolution in a Threespine Stickleback Population.”  Evolution 58(4):814–824.  2004.  Abstract only, read from the Oxford Academic page and checked against PubMed 15154557.  The full text was not obtained.

    1. 26a.

      Section Abstract:  “The 1990 sample was nearly monomorphic for the complete plate morph, which is monomorphic in local sea-run populations; the low plate morph, which is usually monomorphic in local freshwater populations, was absent. By 2001, the frequency of completes had declined to 11%, and lows had increased to 75%.”  “Contemporary evolution in this population indicates that threespine stickleback adapt to freshwater habitats within decades after invasion from the ocean”. ↩

  27. 27.

    Colosimo, Pamela F. et al.  “Widespread Parallel Evolution in Sticklebacks by Repeated Fixation of Ectodysplasin Alleles.”  Science 307(5717):1928–1933.  Mar 25, 2005.  Read from the Schluter lab reprint.  Pages anchored by extracting each PDF page and reading its folio.  Mirror:  Zoology.ubc.ca

    1. 27a.

      Page 1928:  “parallel evolution of stickleback low-plated phenotypes at most freshwater locations around the world has occurred by repeated selection of Eda alleles derived from an ancestral low-plated haplotype that first appeared more than two million years ago.” ↩

  28. 28.

    Payne, D.  “Spread of chloroquine resistance in Plasmodium falciparum.”  Parasitology Today 3(8):241–246.  Aug 1987.  Read in abstract only, from the Europe PMC record.  The full text is behind a paywall and was not obtained.

    1. 28a.

      Section Abstract:  “The drug was introduced in 1934, but was not in large-scale use until the early 1950s.  Anecdotal reports suggest that resistance emerged as early as 1957 both in Colombia and along the then Cambodia-Thailand border area.  But by 1960, resistance in these areas was confirmed - and may represent two separate events.” ↩

  29. 29.

    Wootton, John C. et al.  “Genetic diversity and chloroquine selective sweeps in Plasmodium falciparum.”  Nature 418:320–323.  Jul 18, 2002.  Read in abstract only, from the Europe PMC record.  The full text is behind a paywall and was not obtained.

    1. 29a.

      Section Abstract:  “Using 342 highly polymorphic microsatellite markers from a genetic map, here we show that the level of genetic diversity varies substantially among different regions of the parasite genome, revealing extensive linkage disequilibrium surrounding the key CQR gene pfcrt and at least four CQR founder events.” ↩

  30. 30.

    White, Nicholas J.  “Antimalarial drug resistance.”  Journal of Clinical Investigation 113(8):1084–1092.  Apr 2004.  A review by the Mahidol and Oxford malaria clinician.  Free to read at the journal.  Read from the publisher PDF;  page numbers are the printed folios.  The PDF text flattens exponents, which were restored from context.

    1. 30a.

      Page 1085:  “Resistance to chloroquine in P. falciparum has arisen spontaneously less than ten times in the past fifty years (14).  This suggests that the per-parasite probability of developing resistance de novo is on the order of 1 in 1020 parasite multiplications.”  Reference 14 is Su, Kirkman, Fujioka and Wellems 1997, Cell 91:593–603. ↩

  31. 31.

    Burger, J. et al.  “Absence of the lactase-persistence-associated allele in early Neolithic Europeans.”  Proceedings of the National Academy of Sciences 104(10):3736–3741.  Mar 6, 2007.  Johannes Gutenberg University, Mainz.  Read from PubMed Central;  locators are sections.  Samples from eastern Germany, Hungary, northeastern Poland and northeastern Lithuania, dated archaeologically or by radiocarbon (Table 1).  Mirror:  PMC

    1. 31a.

      Section Abstract:  “We did not observe the allele most commonly associated with LP in Europeans, thus providing evidence for the culture-historical hypothesis, and indicating that LP was rare in early European farmers.” ↩

  32. 32.

    Mathieson, Iain, and Jonathan Terhorst.  “Direct detection of natural selection in Bronze Age Britain.”  Genome Research 32(11–12):2057–2067.  2022.  CC BY-NC 4.0.  Read from the library PDF;  PDF page N is printed page 2056 + N, checked against the folios on PDF pages 6 and 8.  Mirror:  PMC

    1. 32a.

      Page 2063:  “We find that selection for the persistence allele was strongest (s ≈ 0.08) from 150 to 100 generations before the present (roughly 4500–3000 BP) before decreasing to around 0.02 in the past 50 generations. This large change in strength of selection might explain the wide range of estimates from models that assume a constant value”.  Table 1:  “s̄ is the average selection coefficient over the observed time period.”  Table 1 gives LCT s̄ = 0.064;  the bar over the s is lost in the PDF text layer.  The model is genic:  "the a allele has relative fitness 1 + st/2 relative to A" (p. 2057). ↩

  33. 33.

    Jeong, Choongwon et al.  “Long-term genetic stability and a high-altitude East Asian origin for the peoples of the high valleys of the Himalayan arc.”  PNAS 113(27):7485–7490.  2016.  12 authors.  Read from the PMC HTML fetched with bin/pmcget.py.  Copyright.  Mirror:  PMC

    1. 33a.

      Section High-Altitude Functional Alleles:  “In contrast, reads supporting derived alleles at the EPAS1 SNPs were found in two of the three later Samdzong individuals (S35 and S41), but not in the earlier Chokhopani (C1) or Mebrak (M63) individuals.” ↩

  34. 34.

    Huerta-Sánchez, Emilia et al.  “Altitude adaptation in Tibetans caused by introgression of Denisovan-like DNA.”  Nature 512:194–197.  2014.  27 authors.  Read from the PMC author manuscript in the source library, whose title reads "in Tibet" where the published title reads "in Tibetans".  Locators are section headings of that manuscript, not Nature pages.  Mirror:  PMC

    1. 34a.

      Section Abstract:  “Re-sequencing the region around EPAS1 in 40 Tibetan and 40 Han individuals, we find that this gene has a highly unusual haplotype structure that can only be convincingly explained by introgression of DNA from Denisovans or Denisovan-related individuals into humans. Scanning a larger set of worldwide populations, we find that the selected haplotype is only found in Denisovans and in Tibetans, and at very low frequency among Han Chinese.” ↩

  35. 35.

    Dawkins, Richard.  The Greatest Show on Earth: The Evidence for Evolution.  Free Press.  2009.  Copyright.  Print-disabled archive.org copy read with bin/iapdsearch.py;  not stored.  Leaf 378 = p. 364 by the item's page map, confirmed by the book's index entry "vas deferens, 364-5".  Mirror:  Archive.org

    1. 35a.

      Chapter 5, page 116:  “Now, the fascinating thing is that, although caecal valves don’t normally occur in Podarcis sicula and are rare in the family to which it belongs, those valves have actually started to evolve in the population of P. sicula on Pod Mrcaru, the population that has, for only the past thirty-seven years, been evolving towards herbivory.”  “I should repeat that the only thing that is really exceptional about this whole story, and the reason I am telling it here, is that it all happened so extremely rapidly, in a matter of a few decades: evolution before our very eyes.” ↩

  36. 36.

    Herrel, Anthony et al.  “Rapid large-scale evolutionary divergence in morphology and performance associated with exploitation of a different dietary resource.”  Proceedings of the National Academy of Sciences 105(12):4792–4795.  Mar 25, 2008.  Read from the publisher PDF in the Wayback Machine.  Page numbers checked against the printed footers.  Mirror:  PMC

    1. 36a.

      Page 4794:  “Although the presence of cecal valves and large heads in hatchlings and juveniles suggests a genetic basis for these differences, further studies investigating the potential role of phenotypic plasticity and/or maternal effects in the divergence between populations are needed.” ↩

  37. 37.

    Vervust, Bart et al.  “Anatomical and physiological changes associated with a recent dietary shift in the lizard Podarcis sicula.”  Physiological and Biochemical Zoology 83(4):632–642.  2010.  Read online at JSTOR on a free account.  Mirror:  Jstor.org

    1. 37a.

      Page 635:  “Lizards maintained on an arthropod diet showed a reduction in the weight of the small intestine and the hindgut. In addition, we found no evidence of cecal valves in any of the twenty specimens that had been fed arthropods. This was in sharp contrast to the high proportion (12/16) of individuals with valves among the freshly caught individuals”. ↩

  38. 38.

    Sabolić, I. et al.  “Plastic and genomic change of a newly established lizard population following a founder event.”  Molecular Ecology 33(2):e17255.  2024.  Abstract only, read from Europe PMC.  The full text is paywalled, and the author's 2021 University of Zagreb thesis on the same work was not reachable past the repository's robot check.

    1. 38a.

      Section Abstract:  “Diverged traits related to head size and shape or bite force showed moderate heritability in a crossing experiment, but between-population differences in these traits did not persist in a common garden environment. Our results confirm the existence of sufficient additive genetic variance for traits to evolve under selection while also demonstrating that phenotypic plasticity and/or genotype by environment interactions are the main drivers of population differentiation at this early evolutionary stage.” ↩

  39. 39.

    Answers in Genesis.  “Island evolves lizards.”  Answers in Genesis, News to Know.  Jun 6, 2009.  Retrieved Sep 29, 2026.  Unsigned.  The page date is 2009;  the story it answers ran in April 2008.

    1. 39a.

      “But Answers in Genesis anatomist David Menton , after reviewing the research, noted that the original lizards did have the ability to digest plant material; they simply preferred insects for roughly 95 percent of their dietary needs.”  “Furthermore, Menton adds, “The ‘new’ muscular valve they found between the small and large intestine is simply an enlargement of muscles already present in the gut wall at this juncture.””  “Menton also suggests that if the lizards were returned to their original habitat, the cecal valve feature may dwindle as the lizards returned to an insectivorous diet.” ↩

  40. 40.

    Zhuang, Xuan et al.  “Molecular mechanism and history of non-sense to sense evolution of antifreeze glycoprotein gene in northern gadids.”  Proceedings of the National Academy of Sciences 116(10):4400–4405.  2019.  CC BY-NC-ND.

    1. 40a.

      Page 4400:  “We constructed genomic DNA libraries for AFGP-bearing and AFGP-lacking species across the gadid phylogeny and performed fine-scale comparative analyses of the AFGP genomic loci and homologs.”  Page 4401:  “reconstruct the gadid AFGP evolutionary path.”  The first quotation is from the abstract. ↩

  41. 41.

    Chen, Liangbiao et al.  “Evolution of antifreeze glycoprotein gene from a trypsinogen gene in Antarctic notothenioid fish.”  Proceedings of the National Academy of Sciences 94(8):3811–3816.  1997.  Copyright.  Mirror:  PMC

    1. 41a.

      Section Abstract:  “The small sequence divergence (4–7%) between notothenioid AFGP and trypsinogen genes indicates that the transformation of the proteinase gene into the novel ice-binding protein gene occurred quite recently, about 5–14 million years ago”. ↩

  42. 42.

    Near, Thomas J. et al.  “Ancient climate change, antifreeze, and the evolutionary diversification of Antarctic fishes.”  Proceedings of the National Academy of Sciences 109(9):3434–3439.  2012.  Fossil-calibrated family tree of 83 notothenioid species, with the presence or absence of antifreeze in 47 of them mapped onto it to date the origin.  The date comes from the tree, not from the antifreeze gene.

    1. 42a.

      Page 3434:  “we demonstrate that the origin of AFGP occurred between 42 and 22 Ma, which includes a period of global cooling approximately 35 Ma.” ↩

  43. 43.

    Wu, Dong-Dong et al.  “De Novo Origin of Human Protein-Coding Genes.”  PLoS Genetics 7(11):e1002379.  2011.  Library PDF.  CC BY.  Mirror:  PMC

    1. 43a.

      Page 3:  “Since the chimpanzees and humans shared a common ancestor”.  The sentence continues with a divergence time of 5 to 6 million years and a rate of 9.83 to 11.8 genes per million years;  the text layer turns each "about" sign into a comma, so the numbers are given in the prose. ↩

  44. 44.

    Weisman, Caroline M. et al.  “Many, but not all, lineage-specific genes can be explained by homology detection failure.”  PLoS Biology 18(11):e3000862.  2020.  Library PDF.  CC BY.  Mirror:  PMC ↩

  45. 45.

    Byrne, Katharine, and Richard A. Nichols.  “Culex pipiens in London Underground tunnels: differentiation between surface and subterranean populations.”  Heredity 82:7–15.  1999.  Free to read.  Copyright.

    1. 45a.

      Page 13:  “no offspring were recorded from crosses between Underground and surface populations.”  Page 14 of the paper notes that infection with the bacterium Wolbachia can make crosses between mosquito populations fail. ↩

    2. 45b.

      Page 14:  “As it appears plausible that colonization of the London Underground has occurred only once, the reproductive isolation could have accumulated in the few hundred generations since then.” ↩

  46. 46.

    Wieland, Carl.  “Brisk biters.”  Creation 21(2):41.  Mar 1999.  Retrieved Sep 27, 2026.  Read from the web copy, dated 25 January 2006.  Wieland was then the head of the organisation that became Creation Ministries International.  Copyright.

    1. 46a.

      Paragraph 1–2:  “About 100 years ago, bird-biting mosquitoes called Culex pipiens entered the tunnels then being dug for the London Underground (the ‘Tube’).”  “British scientists have now found that it is almost impossible to mate those in the Tube with the ones still living above ground, thus suggesting that they have become a new species”. ↩

  47. 47.

    Haba, Yuki et al.  “Ancient origin of an urban underground mosquito.”  Science 390.  Oct 23, 2025.  Article number eady4515.  Read from the copy hosted by the Royal Museum for Central Africa.  The article is paged 1 to 13, after an unnumbered summary page.  A preprint was posted on bioRxiv on 28 January 2025.  Copyright.  Mirror:  Africamuseum.be

    1. 47a.

      Page 1:  “Population genomic analysis of ~350 contemporary and historical samples counters the popular hypothesis that molestus originated belowground in London <200 years ago. Instead, we show that molestus first adapted to human environments aboveground in the Mediterranean or Middle East over the course of more than 1000 years, possibly in association with ancient agricultural civilizations of the Middle East.” ↩

    2. 47b.

      Section Research article summary, Conclusion:  “Our work debunks one of the most widely cited examples of rapid urban adaptation—an example that has captured the attention of scientists and laypeople for 25 years.”  The summary page is not numbered and comes before page 1. ↩

    3. 47c.

      Section Research article summary, Introduction:  “Form molestus features prominently in textbooks as well as scholarly reviews of urban adaptation. Yet, the hypothesis of in situ urban evolution has never been rigorously tested.” ↩

    4. 47d.

      Page 4:  “Accurate assignment of dates to this rCC curve requires knowledge of the de novo mutation rate (μ) and generation time (g), neither of which has been directly measured for Cx. pipiens in nature.”  “suggest that peak rates of divergence occurred ~2000 years ago (Fig. 4C, dashed arrow), whereas minimum and maximum reasonable values (32) lead to split times anywhere between 1300 and 12,500 years ago”.  The rate used is 4.85 × 10-9 per site per generation and the generation time is 20 days.  The exponent was checked against the Methods on page 9. ↩

  48. 48.

    Ohno, Susumu.  “Birth of a unique enzyme from an alternative reading frame of the preexisted, internally repetitious coding sequence.”  Proceedings of the National Academy of Sciences USA 81(8):2421–2425.  Apr 1984.  Contributed by Ohno as a member of the Academy, 4 January 1984.  Read from the scanned pages.  Pages are the printed folios in the running heads.  Copyright.  Mirror:  Archive.org

    1. 48a.

      Page 2423, figure 1:  “I assume that the longer open reading frame identified as PR.C. was the original coding sequence of this stretch of plasmid DNA until several decades ago.”  Figure caption. ↩

    2. 48b.

      Page 2424:  “It should be further noted that if T marked by an arrow in the 3rd row of Fig. 1a is deleted, this longer open reading frame identified as PR.C. (the abbreviation of the preexisted coding sequence) in Fig. 1 a and b starts from the first A-T-G in 1st row of Fig. 1a and ends in T-G-A that are the last three bases in the last row of Fig. 1b, thus specifying a polypeptide chain 427 amino acid residues long.  I propose that this was the original coding sequence contained in the 1295-base-long stretch of pOAD2 plasmid harbored by Flavobacterium Sp. KI72.”  Left column.  The text layer of the scan reads "Fig. la" and "K172";  the page image prints "Fig. 1a" and "KI72". ↩

  49. 49.

    Thwaites, William M.  “New Proteins Without God's Help.”  Creation/Evolution Journal 5(2):1–3.  1985.  Issue 16, Summer 1985, published by the National Center for Science Education.  Read from the NCSE web copy, which marks the print page breaks and says it "might differ slightly from the print publication".  Copyright.

    1. 49a.

      Page 2:  “Now it has happened!  Not one, but two, new proteins have been discovered.  In all probability new proteins are forming by this process all the time, but this seems to be the first documentation of this phenomenon.”  “Since nylon first came into commercial production in 1940, we know that the new enzymes have formed since that time.” ↩

  50. 50.

    Kato, Ko et al.  “Amino acid alterations essential for increasing the catalytic activity of the nylon-oligomer-degradation enzyme of Flavobacterium sp.”  European Journal of Biochemistry 200(1):165–169.  1991.  Free to read at the publisher.  EII is NylB and EII' is NylB'.  Copyright.

    1. 50a.

      Page 168:  “Our studies demonstrated that among the 47 amino acids altered between the EII and EII' proteins, a single amino acid substitution at position 181 was essential for the activity of 6-aminohexanoate-dimer hydrolase and substitution at position 266 enhanced the effect.  The remaining 45 amino acid substitutions had no effect on the activity, although these 45 alterations included many non-synonymous substitutions”.  “This indicates that only two amino acid replacements in the EII' protein are enough to increase the activity by 1000-fold.” ↩

  51. 51.

    Negoro, Seiji et al.  “X-ray crystallographic analysis of 6-aminohexanoate-dimer hydrolase:  molecular basis for the birth of a nylon oligomer-degrading enzyme.”  Journal of Biological Chemistry 280(47):39644–39652.  Nov 25, 2005.  The structure is of Hyb-24, a laboratory hybrid of EII and EII' with activity at the level of EII', because EII itself would not crystallise.  Read from the archived publisher PDF.  Copyright.  Mirror:  Archive.org

    1. 51a.

      Page 39652:  “These results indicate that the G181D and H266N are amino acid alterations specific for the increase of nylon oligomer hydrolysis.  Thus, the nylon oligomer-degrading enzyme (EII) is considered to have evolved from preexisting esterases with β-lactamase folds.” ↩

  52. 52.

    Boxhorn, Joseph.  “Observed Instances of Speciation.”  TalkOrigins Archive.  Sep 1, 1995.  Retrieved Sep 27, 2026.  The archived copy prints “Last Update: September 1, 1995”.  The byline is blank in both copies read;  the attribution to Boxhorn is the archive's.  Copyright.  Mirror:  Archive.org

    1. 52a.

      Section 5.5.1 Apple Maggot Fly:  “It is important to note that some of the leading researchers on this question are urging caution in interpreting it.” ↩

  53. 53.

    Axe, Douglas D.  “Estimating the Prevalence of Protein Sequences Adopting Functional Enzyme Folds.”  Journal of Molecular Biology 341(5):1295–1315.  2004.  Axe was then at the Babraham Institute, Cambridge.  Library PDF.  The text layer drops the superscript of 1077.  Copyright.

    1. 53a.

      Page 1295:  “the overall prevalence of sequences performing a specific function by any domain-sized fold may be as low as 1 in 1077”. ↩

    2. 53b.

      Page 1298:  “the approach should be first to obtain an extensively degraded reference sequence that just passes a low selection threshold”. ↩

  54. 54.

    Miller, Brian, and Casey Luskin.  “Defending Douglas Axe on the Rarity of Protein Folds.”  Science and Culture Today, Discovery Institute.  Nov 29, 2023.  Retrieved Aug 2, 2026.  Both authors work for the Discovery Institute.  The page's text layer prints the exponent as "10 77";  the superscript is restored here.  Copyright.

    1. 54a.

      Paragraph 1:  “found that the likelihood of a chance sequence of 153 amino acids generating the stable, functional fold needed for the larger domain in that enzyme was as low as 1 in 1077.”  Section Scientific Objections to Axe’s Research:  “Axe (2000, 2004)’s work on the larger beta-lactamase protein domain indicates that only 1 in 1077 sequences are functional”. ↩

  55. 55.

    Dembski, William A.  No Free Lunch: Why Specified Complexity Cannot Be Purchased without Intelligence.  Rowman and Littlefield.  2002.  Read on archive.org by paragraph search.  The item cannot be borrowed and was not stored.  The printed page is the leaf minus 32, anchored at leaves 60 and 400 from the site's page map.

    1. 55a.

      Page 293:  “Randomly picking 250 proteins and having them all fall among those 500 therefore has probability (500/4,289)^250, which has order of magnitude 10^-234 and falls considerably below the universal probability bound”.  The OCR loses both exponents.  They are recovered by recomputing;  see the note on the arithmetic. ↩

  56. 56.

    Dembski, William A.  Specification: The Pattern That Signifies Intelligence.  Aug 15, 2005.  Version 1.22 of the author's manuscript, 41 pages, numbered at the foot.  Page numbers here are the manuscript's.  Later printed in Philosophia Christi 7(2):299–343.

    1. 56a.

      Page 25:  “is an evolutionary chance hypothesis that takes into account Darwinian and other material mechanisms”.  “The precise calculation of P(T|H) has yet to be done.” ↩

  57. 57.

    Dembski, William A.  “Comment on "Some Thanks for Professor Olofsson".”  Uncommon Descent.  Dec 3, 2008.  Posted 6:24 PM PDT under a post by DaveScot of 25 November 2008.  On the archived page each author's name follows the comment.

    1. 57a.

      Paragraph point 2:  “The challenge for determining whether a biological structure exhibits CSI is to find one that's simple enough on which the probability calculation can be convincingly performed but complex enough so that it does indeed exhibit CSI. The example in NFL ch. 5 doesn't fit the bill.” ↩

  58. 58.

    Behe, Michael J.  Darwin's Black Box: The Biochemical Challenge to Evolution.  Free Press.  10th anniversary edition.  2006.

    1. 58a.

      Page 40:  “Even if a system is irreducibly complex (and thus cannot have been produced directly), however, one can not definitively rule out the possibility of an indirect, circuitous route. As the complexity of an interacting system increases, though, the likelihood of such an indirect route drops precipitously.” ↩

    2. 58b.

      Page 95:  “by Doolittle's own account no blood clotting appears until at least the third step.”  Page 96:  “The bottom line is that clusters of proteins have to be inserted all at once into the cascade.”  Page 97:  “The fact is, no one on earth has the vaguest idea how the coagulation cascade came to be.” ↩

  59. 59.

    Doolittle, Russell F.  “Step-by-step evolution of vertebrate blood coagulation.”  Cold Spring Harbor Symposia on Quantitative Biology 74:35–40.  2009.  Publisher PDF read through a browser session;  sheet N is printed page 34 + N.  Copyright.

    1. 59a.

      Page 38:  “The total absence of clotting factors in basal chordates and the existence of a relatively complex system in lampreys indicate that it is not likely we will find any extant creatures with truly rudimentary clotting systems, leaving us to speculate on how the system might have gotten started.” ↩

  60. 60.

    Sarfati, Jonathan.  “Arguments we think creationists should NOT use.”  Creation Ministries International.  Feb 7, 2006.  Retrieved Sep 27, 2026.  The page carries the date 7 February 2006 and has been revised since.  The list was first written when Creation Ministries International and Answers in Genesis were one organisation.  Copyright.

    1. 60a.

      Section “There are no beneficial mutations.”:  “This is not true, since some changes do confer an advantage in some situations. Rather, we should say, ‘We have yet to find a mutation that increases genetic information in ways needed for microbes-to-man evolution to be feasible.”  An entry under the heading “Which arguments should definitely not be used?” ↩

  61. 61.

    Purdom, Georgia.  “Antibiotic Resistance of Bacteria: An Example of Evolution in Action?”  Answers 2(3).  Jul 10, 2007.  Purdom holds a PhD in molecular genetics and works for Answers in Genesis.  From Answers magazine, July to September 2007.  Copyright.

    1. 61a.

      Section What Does It Really Prove?:  “Mutation and natural selection, thought to be the driving forces of evolution , only lead to a loss of functional systems.”  The space before the comma is in the page's text, where a link ends. ↩

  62. 62.

    Batten, Don.  “Antibiotic resistance: Evolution in action?”  Creation 39(4):46–48.  Oct 2017.  First appeared in Creation 39(4), October 2017; the web version was posted 11 March 2019 and updated 27 March 2019. Cited by its print appearance, which is the line creation.com prints at the head of the article.

    1. 62a.

      Section Inactivation of the antibiotic:  “It might appear that a bacterium faced with an antibiotic and then acquiring such an ability would be a great example of evolution—a new enzyme; a new gene.  However, in the instances known, the machinery to manufacture the antibiotic-destroying enzyme has never been observed to arise by mutation.” ↩

  63. 63.

    Wood, Todd Charles.  “Perspectives on AGEing, a young-earth creation diversification model.”  Proceedings of the Fifth International Conference on Creationism, 479–489.  2003.  Copyright.

    1. 63a.

      Page 479:  “Since we see no ‘speciation’ like this today or in historical times, we can infer that the mechanism of intrabaraminic diversification is qualitatively different from modern speciation”. ↩

    2. 63b.

      Page 479:  “This would require the origin of 150 species of horse prior to the gift of donkeys (Equus asinus) from Pharaoh to Abram (Gen. 12:16), which took place ~400 years after the Flood.” ↩

  64. 64.

    Sanford, John C.  Genetic Entropy & the Mystery of the Genome.  FMS Publications.  2nd edition.  2005.  Sanford is a plant geneticist, formerly at Cornell.  Library PDF, laid out two printed pages to a sheet;  PDF page 40 holds printed pages 64 and 65.  Copyright.

    1. 64a.

      Page 65, figure 4:  “Dr. Crow (1997) indicates that he believes the fitness of the human race is presently degenerating at 1-2% per generation, due to the accumulation of mutations. A 1% decline in fitness per generation (beginning with a fitness of 1) is shown for a hypothetical human population - over a period of 300 generations (6,000 years).” ↩

  65. 65.

    Crow, James F.  “The high spontaneous mutation rate: Is it a health risk?”  Proceedings of the National Academy of Sciences 94(16):8380–8386.  1997.  Crow was a population geneticist at the University of Wisconsin.  Read from the PubMed Central copy, which has no page numbers.  Copyright.  Mirror:  PMC

    1. 65a.

      Section The Current Human Population:  “If we are like Drosophila, the decrease in viability from mutation accumulation is some 1 or 2% per generation. This is more than compensated for by much more rapid environmental improvements”. ↩

  66. 66.

    “Proceedings of the Microbe Forum, June 2007.”  Answers Research Journal 1:1–6.  Edited by Georgia Purdom and Joseph W. Francis.  2008.  Joseph Francis's contribution, "Creation Microbiology and the Origin of Disease", is on p. 3.  Library PDF.  Copyright.

    1. 66a.

      Page 3:  “From a creation view, it appears, then, that the origin of microbial based disease has at least two primary causes, (1) post-Fall genetic alteration of the original good microbe and/or (2) post-Fall displacement or movement of the microbe from the site where it performed its beneficial function.” ↩