Living Fossils

"The so-called living fossils that people like to cite as examples of evolutionary stasis are largely a myth." — Thomas J. M. Schopf, University of Chicago, as reported from the 1980 Field Museum meeting¹

"I cannot be 100 percent certain that any one comparison presented in this book (of a dinosaur-era fossil and a modern living organism) is, in fact, the same genus or species." — Carl Werner, Living Fossils, 2008²

Summary

A living fossil is a living organism that looks like fossils found deep in the rock record.  Charles Darwin coined the phrase in 1859.  Young-earth creationists cite these organisms as evidence that life has not evolved, and as evidence that the rocks holding their fossils are not hundreds of millions of years old.  Mainstream biologists answer that the label is loose, that the famous cases have in fact changed, and that evolution never predicted constant change.

The question only has an answer once the taxonomic rank is named.  "The horseshoe crab has not changed in 450 million years" can mean the species, the genus, the family or the order, and the answer is different at each.  Measured that way, the evidence gives this picture.

Eight conclusions, stated first.

1. At family rank and above, the creationist observation is real, and mainstream specialists say so in print. Coelacanths, horseshoe crabs, nautiloids, lingulid brachiopods and ginkgos each keep a recognisable body plan through most of the rock record. A 2014 coelacanth study calls them "almost living fossils."³ A 2014 horseshoe crab study coins a new word, "stabilomorph," for genera that barely change.⁴ A 2024 analysis of 82 coelacanth species finds that the coelacanth's discrete anatomical characters "virtually stopped evolving after the Cretaceous."⁵
2. At genus rank, most of the famous cases fail. No fossil of the living coelacanth genus Latimeria, the Wollemi pine Wollemia or the monoplacophoran Neopilina is known. The Jurassic horseshoe crab once called Limulus has been moved to another genus, which leaves Limulus with a Cretaceous-to-Recent record.⁶ No Paleozoic or Mesozoic species is assignable to Lingula in the strict sense.⁷ The tuatara genus Sphenodon first appears, tentatively, in Miocene rock. Nautilus first appears in Eocene to Oligocene rock. The half-billion-year spans quoted for these animals belong to their families, orders and classes.
3. At species rank, only ginkgo crosses into Mesozoic rock. Royer, Hickey and Wing, writing in Paleobiology, call the fossil leaf species Ginkgo adiantoides "indistinguishable from modern G. biloba" and say their own ecological results support treating the two as one species.⁸ G. adiantoides first appears in Early Cretaceous rock, so on that reading the living ginkgo species, judged by its leaves, reaches back into the Cretaceous.⁸ It is the best single case the creationist argument has. No other organism on the usual list has a living species in rock older than the Pliocene.
4. Thomas Schopf's 1980 objection is right at species rank and misses the creationist claim. Schopf argued that soft parts can change while shells stay the same, so a living fossil's stasis is partly an illusion.¹ The genomes now test this, and the results split: the coelacanth's protein-coding genes evolve unusually slowly,⁹ while Lingula's gene families turn over faster than any other bilaterian's sampled.¹⁰ The two internal organs that have fossilized also split. A Carboniferous horseshoe crab's nervous system has the same basic layout as a living one's,¹¹ while the living coelacanth's lung is a vestige of the large, plated lung its Mesozoic relatives carried.¹² But young-earth creationists claim stasis of the created "kind," usually near family rank, and Schopf's argument does not reach that rank.
5. Long stasis costs a young earth almost nothing and costs deep time a mechanism. At Gingerich's measured upper rate of change, used here only for scale, a trait that wandered freely for the roughly one million generations the conventional timescale gives the coelacanth since the Cretaceous would move about a hundred standard deviations. In the few thousand years a young-earth chronology allows, it would move about one. So deep time needs active constraint, such as stabilizing selection, to explain stasis, and the bound on that constraint is still unexplained.¹³ The mainstream's own measurements say the same without the illustration: Michael Lynch found in 1990 that skull shape in almost every mammal lineage he measured had diverged "substantially below the minimum neutral expectation," the change that mutation and random drift alone would produce, and put the shortfall down to stabilizing selection.¹⁴ But the young-earth side cannot treat stasis as a prediction either, because its own models need fast diversification within kinds after the Flood.
6. Neither model predicts which lineages stall. A creationist statement of what the Flood model predicts lists three outcomes, changed within the kind, extinct, or exactly the same, and every fossil fits one of them.¹⁵ The mainstream account explains stasis after the fact with selection and habitat. Stasis therefore does not decide between the two models.
7. The gaps are real rarity, not missed sampling, and that binds both sides. At the rate coelacanths appear in Late Cretaceous marine fish beds, about 54 of the 1,686 Cenozoic marine collections holding ray-finned fishes should contain a coelacanth. None does. The same test predicts about 156 Jurassic-to-Pliocene monoplacophoran collections where there are none, and about 28 Eocene-to-Pliocene tuatara-relative collections where there are two.¹⁶ On the mainstream timescale these lineages hid in refuges for tens of millions of years. On a young-earth timescale they lived through the later Flood or the post-Flood centuries where the rocks did not sample them. Neither side can treat the absence as a sampling accident.
8. Most living genera have no old fossils, which cuts against "modern animals in dinosaur rock." Of 1,351 living mammal genera, none has a fossil in Mesozoic rock in the Paleobiology Database; among about 5,400 living genera of mammals, birds, lizards, snakes and amphibians, fewer than ten do, and those few are doubtful single records. Among marine invertebrates it is common: about one living bivalve genus in six has a Mesozoic or older record.¹⁷ The zero for mammals does not count against the young-earth claim at the rank of the kind, because young-earth models also expect land animals to have split into new genera after the Flood.¹⁸

Verdict, in two sentences.  Living fossils show real conservatism at family rank, which young-earth creationists correctly say mainstream textbooks underplay, and the ginkgo is a genuine species-level case;  but "unchanged for hundreds of millions of years" almost always slides from a living species or genus to a fossil family or order.  Long stasis fits a short timescale more cheaply than a long one, yet it does not discriminate between the models, and the gaps between appearances require each side to place the survivors in habitats the rocks did not sample.

This article builds on two others.  Stasis and punctuated equilibrium in general are argued in Trends in the Fossil Record, Part 1.  Why Lazarus taxa do not disturb the order of the fossil record, and why "absence of evidence" cuts both ways, is argued in The Order of the Fossil Record, Part 3.  Insects in amber are treated in Amber and the Flood.

Part 1: What the phrase claims, and at which rank

Darwin's phrase was about survival, not about standing still

Darwin used the phrase twice in 1859.  Writing of the platypus and the South American lungfish, he said:  "These anomalous forms may almost be called living fossils;  they have endured to the present day, from having inhabited a confined area, and from having thus been exposed to less severe competition."¹⁹  In his conclusion he called such groups forms "which may fancifully be called living fossils."¹⁹  He was explaining why old lineages survive.  He was not claiming that they had not changed.

The modern literature has no single definition.  Niles Eldredge and Steven Stanley's 1984 casebook set one rule:  "a living species must be anatomically very similar (bordering on identity) to a fossil species which occurs very early in the history of the lineage."²⁰  Thomas Schopf listed six rival definitions in the same year, from "a living species that has persisted over a very long interval of geologic time" to "a living species that was once thought to be extinct."²⁰  Christian Emig, in his paper on Lingula, ruled that "only a species can be referred to as a living-fossil - supra-specific taxa are excluded automatically."⁷  Scott Lidgard and Alan Love reviewed the field in 2018 and found the phrase "now used to describe entities ranging from viruses to higher taxa, despite recent warnings of misleading inferences."²¹

What the young-earth side claims

Carl Werner's Living Fossils (2008) is the fullest creationist treatment.  Werner set out to test evolution with a stated prediction:  "if evolution was not true, then animals and plants would not change significantly over time.  Accordingly, I predicted I should find fossils of modern animal and plant species in the 'older' fossil layers."²  After visiting museums and dig sites, he wrote that if he would "ignore the genus and species names and simply compare the fossils found in dinosaur rock layers to modern forms, I see a lack of significant change in all of the major animal phyla and all of the major plant divisions."²  He was candid about the limits.  He had little access to the specimens and never examined one under a microscope, so "I cannot be 100 percent certain that any one comparison presented in this book (of a dinosaur-era fossil and a modern living organism) is, in fact, the same genus or species."²

Creation Ministries International states the theological reading plainly.  Creatures were "created to be fruitful and multiply after their kind.  So the fact that modern creatures have 'stayed the same' as their fossilised ancestors is no surprise at all."²²  Don Batten adds the argument against the other side:  "Evolution is about change, and putting 'evolutionary' in front of 'stasis', does not explain the stasis in terms of evolution."²³

The young-earth claim is about the created kind, and the kind is usually placed near family rank.  Nathaniel Jeanson and Jason Lisle write that the kinds of Genesis 1 "appear to be best approximated by the taxonomic rank of family, not species."²⁴  Jake Hebert, discussing coelacanths in the Journal of Creation, says the fossil genera "may all belong to a few Genesis kinds, perhaps even one."²⁵  Werner's own method sets genus and species names aside.  So a creationist who says a coelacanth is "unchanged" is usually claiming that it is still a coelacanth, not that it is the same species as a Cretaceous fossil.  The arguments below hold the claim to that rank, and do not attack a species-level claim the creationist sources do not make.

The rank rule

Every "unchanged for N million years" statement in this article is tested at four ranks:  species, genus, family, and order or class.  A statement passes at a rank when fossils assigned to that living taxon by specialists occur in rock of that age.  It fails when the old fossils belong to other genera or families.  The difference matters because the spans are very different.  The Paleobiology Database places the oldest coelacanth (subclass Actinistia) in the Early Devonian, the oldest member of the living coelacanth family (Latimeriidae) in the Early Triassic, and no fossil at all in the living genus Latimeria.²⁶

How often a living genus turns up in old rock

Werner's thesis is that modern organisms are found in "dinosaur rock."  At genus rank this can be counted.  The Paleobiology Database lists, for each living genus it knows, its oldest fossil occurrence.  Dividing by the number of living genera in the GBIF taxonomic backbone, which has no paleontological bias, gives the share of living genera with fossils older than a given point.¹⁷

Two things follow.  First, Werner's claim fails at genus rank for land vertebrates.  No living mammal genus, and at most a handful of bird, lizard and amphibian genera, has a fossil in Mesozoic rock in the database, and the few records that exist are single latest-Cretaceous identifications, several of them database errors.  Werner concedes the large-mammal part himself:  "I did not find any large mammals."²⁷  His claim holds only at the higher rank he chose, mammals in general, birds in general.  The zero does not count against the young-earth model as its geologists state it, though.  John Whitmore holds that "there were less than 300 mammal kinds on the ark that diversified into all the mammal species that we have today,"¹⁸ so on his model, too, no living mammal genus should turn up in Flood rock.  The census refutes the genus-level reading of "modern animals in dinosaur rock," not the kind-level claim, and since the mainstream predicts the same zero, it decides nothing between the two.

Second, genus-level persistence is ordinary in the sea.  About one living bivalve genus in six, and about one living shark or ray genus in four, has Mesozoic or older fossils.  A creationist count from the same database reached the same order of magnitude:  John Whitmore, writing for Answers in Genesis, found "99 living fossil genera" from the Paleozoic and "548" from the Mesozoic.¹⁸  By this measure the famous living fossils are not rare exceptions.  They are the best-known members of a large class of long-ranging marine genera.

Part 2: The cases, measured rank by rank

Each case below gives the claim, what survives at each rank, what was measured in the anatomy, and what the genome shows where one has been read.  A summary table closes the part.

The coelacanth

The claim.  An Institute for Creation Research page says the coelacanth caught in 1938 "showed the same well-designed form as the fossils," and asks:  "Where is any evidence of natural selection having made even one significant change in this fish over its supposed 400-million-year existence?"²⁸

By rank.  The 400 million years belong to the subclass.  The Paleobiology Database places the oldest coelacanth in the Early Devonian, the oldest member of the living family Latimeriidae in the Early Triassic, and no fossil in the living genus Latimeria.²⁶  The youngest coelacanth fossil known is an isolated ossified lung from Late Maastrichtian phosphates in Morocco, the last stage of the Cretaceous.²⁹  Peter Forey made the rank point in the 1984 casebook itself:  "Latimeria fails to comply with (1) and (2) and can only comply with (3) if the group is specified," where (3) is "a Recent member of natural group known to have been more diverse in the past."³⁰

Anatomy, measured.  Two measurements agree that the coelacanth line has gained fewer anatomical novelties than its relatives.  Lionel Cavin and Guillaume Guinot counted unique derived characters per million years on the conventional timescale:  about 0.32 for the tetrapod lineage, 0.14 for ray-finned fishes, and 0.04 to 0.05 for coelacanths, so the tetrapod rate is about six times the coelacanth rate.³  They conclude that change "in the actinistian lineage is markedly lower than in the other two clades, at least after the Late Devonian early morphological diversification of the group."³  Alice Clement and colleagues analysed 87 species coded for 322 characters in 2024.  They found that discrete characters, the ones that mark new structures, "virtually stopped evolving after the Cretaceous," so that for those characters Latimeria "has indeed been frozen in evolutionary time."⁵

The same study found the opposite for shape.  Proportions and counts of parts "have continued to evolve at typical Mesozoic rates," and in body shape "Latimeria is positioned away from its closest Mesozoic relatives."⁵  Its authors' summary keeps both halves:  coelacanths "might indeed be considered living fossils due to their lack of major recent innovations, but in more subtle features and their DNA, they have continued to evolve at more normal rates for vertebrates."⁵  Didier Casane and Patrick Laurenti add that extinct coelacanths include eel-like, piranha-like and trout-like forms, and conclude that "extant coelacanths have not remained morphologically static since the Devonian."³¹

Genome.  The 2013 coelacanth genome paper found its proteins "significantly more slowly evolving (0.890 substitutions per site) than the lungfish (1.05 substitutions per site), chicken (1.09 substitutions per site)" and mammals (1.21).⁹  Casane and Laurenti point out that "the four recent molecular studies that explicitly concluded that the genome of Latimeria is slowly evolving also explicitly assumed that coelacanths are 'living fossils'."³¹  A 2024 survey of 1,105 exons across about 480 vertebrate species by Chase Brownstein and colleagues found that "classic living fossil lineages, such as coelacanths and rhynchocephalians, have rates of molecular evolution similar to most other vertebrate clades."³²  The molecular question is open.

Verdict.  At subclass and family rank the coelacanth is a real case of conservatism, and the mainstream says so.  At genus rank there are no fossils to compare.  "The same form" is true of its skeleton's major structures and false of its shape.

The tuatara

By rank.  The order Rhynchocephalia runs from the Early Triassic in the database.  After the Cretaceous its record is thin:  Kawasphenodon peligrensis from the Paleocene of Patagonia, which belongs to "at least one group of rhynchocephalians not related to the extant Sphenodon,"³³ a doubtful record from Paleocene Togo, Early Miocene jaws from New Zealand, and then Pleistocene and Holocene bones.  The Miocene jaws are "closely resembling those of the extant tuatara (Sphenodon)" and bridge "a gap of nearly 70 million years in the rhynchocephalian fossil record" on the conventional timescale.³⁴  The database files them only as "cf. Sphenodon."  So the living genus is not securely known before the Pleistocene.

Anatomy, measured.  Jorge Herrera-Flores, Thomas Stubbs and Michael Benton proposed a test:  "a living fossil should show both statistically significantly slow rates of morphological evolution and it should be morphologically conservative."³⁵  Using the shape of the lower jaw across 31 rhynchocephalians, they found Sphenodon "the seventh most similar to the average shape," with slow rates, and concluded:  "Our study confirms that Sphenodon fits both of these criteria and so is a 'living fossil'."³⁵  Felix Vaux and colleagues replied that "the conclusion of morphological conservatism for tuatara was based on the shapes of one specimen per species," including a single tuatara jaw, although museums hold many.³⁶  The original authors answered in the same journal;  that reply was not obtained for this article.

Genome.  A 2008 study of ancient tuatara DNA reported "the highest rate of molecular change recorded in vertebrates."³⁷  The 2020 tuatara genome, measuring a different thing (substitutions at sites where a change does not alter the protein), found the opposite:  0.00159 substitutions per site per million years against a median of 0.00178 for lizards and snakes, which "makes the tuatara the slowest-evolving lepidosaur yet analysed."³⁸  The same paper's abstract calls the rate "moderate."³⁸  The difference from lizards and snakes is about 11%.

Verdict.  Order-level persistence with a thin record, a living genus with no secure pre-Pleistocene fossil, and anatomical conservatism that rests on one bone per species.

Gars, the best molecular case

Gars are less famous and make a stronger case.  The living genus Lepisosteus has a continuous record in the database from the Early Cretaceous to the Pleistocene, in more than 400 collections.²⁶  Brownstein and colleagues found that "gar exons consistently evolve between 0.5 and 3 orders of magnitude more slowly than any other major vertebrate clade."³²  Their explanation is a mechanism "decoupled from selection that we speculate is linked to a highly effective DNA repair apparatus."³²  Gars are a genuine case of slow evolution at genus rank, measured in both bones and genes, and they are a mainstream result.

Horseshoe crabs

The claim.  The mainstream states this one as strongly as the creationists do.  A 2020 genome paper opens:  "The evolutionary history of horseshoe crabs, spanning approximately 500 million years, is characterized by remarkable morphological stasis and a low species diversity with only four extant species."³⁹  Neither that paper nor an earlier horseshoe crab genome study measures a single anatomical character;  the stasis is a background statement.⁴⁰

By rank.  The order Xiphosura reaches back to the Ordovician.⁶  The living family Limulidae runs in the database from the Mississippian to the Late Eocene, with nothing in the Oligocene to Pleistocene.²⁶  The genus is where the famous claim was made.  In 2014 Adrian Kin and Błażej Błażejowski described Limulus darwini from Late Jurassic rock in Poland and concluded that "the genus Limulus existed about 148 million years ago and has survived to the present day in an almost unchanged form."⁴  Their comparison rested on moulted shells of juveniles, which matched juvenile L. polyphemus in "about 85% of corresponding features."⁴  In 2020 James Lamsdell's phylogenetic revision of the whole order found Limulus polyphyletic, moved the Polish species to the genus Crenatolimulus, and concluded that "Limulus coffini Reeside & Harris, 1952 remains as the only extinct species demonstrably assignable to the genus."⁶  L. coffini comes from the Late Cretaceous Pierre Shale of Colorado.  The living species L. polyphemus has no fossil record, which was Schopf's point in 1980 (Part 3).

Anatomy.  Lamsdell's revision adds that horseshoe crabs "have in the past exhibited a relatively high species diversity … and a greater variation in both morphology and ecology … than their modern representatives."⁶  So the four living species are a narrow survivor of a wider past.

The one study that measured horseshoe crab shape across the whole record found a mosaic.  Russell Bicknell and nine colleagues placed landmarks on 178 specimens of 44 fossil and living species in 21 genera, from the Mississippian to the present.⁴¹  They report "developmental stasis since at least the Pennsylvanian and shape stasis since the Late Jurassic," two bursts of varied shapes in the Carboniferous and Triassic that "coincide with forays into habitation of marginal environments," and a body size that "increased gradually."⁴¹  After the Triassic, they write, horseshoe crabs "converged upon a static form strikingly similar to those from the early Palaeozoic."⁴¹  For the unchanging growth pattern they offer either a developmental constraint or stabilizing selection, and they grant that stabilizing selection "would require variable rates to be strongly non-adaptive and it is unclear why that would be the case."⁴¹  So horseshoe crab shape is measured, not assumed, to be static at order rank from the Jurassic rocks upward, and its size is measured to have grown.

Genome.  Horseshoe crab genomes are not slow or simple.  They show whole-genome duplication, probably three rounds in the mangrove horseshoe crab,³⁹ and the authors of the first study note that this "suggests that genomic diversity need not always be" reflected in the body.⁴⁰

Verdict.  The horseshoe crab body plan is conserved at order rank, and its shape is measured to be static from the Jurassic rocks upward while its size grew;  the living genus is known with confidence only from the Late Cretaceous;  the living species has no fossils.

The nautilus

By rank.  Popular accounts give the nautilus 500 million years.  That span belongs to nautiloid cephalopods as a whole, most of which had straight or loosely coiled shells.  The coiled order Nautilida runs from the Ordovician in the database, and its families changed greatly through time.²⁶  The genus Nautilus has 110 occurrences in the database, starting in the Pennsylvanian, but the Paleozoic and Triassic ones are old uses of the name, several printed with quotation marks or question marks.²⁶  A 1996 restudy of Nautilus praepompilius from the Chegan Formation of Kazakhstan "confirms that this species is assignable to Nautilus, and that this extant genus extends back to the late Eocene."⁴²  The database adds one Paleocene record of that species from Australia, and the record of the genus stays sparse until the Pleistocene.  The living species N. pompilius is known as a fossil from one fragmentary shell in Early Pleistocene deep-water sediments of the Philippines, which its describers refer to the species only "with reservation" and call "the first and oldest fossil record of N. pompilius."⁴³

Genome.  A study of genome-wide markers across the living nautiluses calls them "classic examples of true living fossils," and then finds that "with the exception of Nautilus macromphalus, all previously described species are at odds with genomewide data, testifying to the prevalence of cryptic species among living fossils."⁴⁴  Shells that look alike hide species that are genetically distinct, which is Schopf's point again.

Verdict.  Conservatism of the coiled shell at order rank, a young genus in the record, and a living species known only from the Pleistocene.

Lingula

The claim.  Lingula is often called the oldest living animal genus, on the strength of tongue-shaped brachiopod shells in Cambrian rock.  The database still holds 1,046 occurrences filed under "Lingula," the oldest from the Middle Cambrian of China.²⁶

By rank.  The brachiopod specialist Christian Emig re-examined the family in 2003.  He found that "the living-fossil argument rests on a single piece of evidence;  the typical shape of the shell," and that "no Palaeozoic or Mesozoic species is assignable to Lingula sensu stricto."⁷  In his reading, after the Devonian the family is represented "from the Carboniferous to the Cretaceous only by Lingularia, and since the Tertiary, or possibly from the Late Cretaceous, by the two extant genera Lingula and Glottidia."⁷  The half-billion-year span belongs to the shell shape of the superfamily, not to the genus.

Genome.  The Lingula genome surprised its authors:  "contrary to its reputation as a 'living fossil,' the Lingula genome has been actively evolving, with a disorganized Hox cluster and recently expanded gene families."¹⁰  Its gene families turn over at a rate "the highest among bilaterians" in their comparison.¹⁰  The same paper concedes the morphological point:  Silurian lingulids have "morphology very similar to that of extant species."¹⁰  Its basic metabolic genes are the slowest-evolving in its comparison set.

Verdict.  A conserved shell at family and superfamily rank, a genus that is not securely older than the latest Cretaceous, and a genome that is not static.

Monoplacophorans

The claim.  Limpet-like monoplacophorans were known only from Cambrian to Devonian rock until living specimens were dredged from abyssal depths off Costa Rica in the 1950s.  David Lindberg's review grants the strongest version of the stasis claim, at class rank:  the living forms "have retained the simple limpet shell morphology with the same shell microstructure;  the shell musculature appears unchanged over a half a billion years."⁴⁵  He also notes that the stasis is inferred:  "it is assumed that they have been in morphological and/or physiological stasis and therefore retain primitive characters for their group."⁴⁵

By rank.  The living family Neopilinidae, and the genus Neopilina, have no fossils in the database.  The gap is shorter than the "almost 400 million years" usually quoted.  The database's class Tergomya, which holds most fossil monoplacophorans, has records into the Permian and possibly the Late Triassic, and in 1990 a few shells of the living species Micropilina minuta were found in Pleistocene clays in Calabria, southern Italy, "conchologically indistinguishable" from the modern animal.⁴⁶

Genome.  The first molecular dating of living monoplacophorans, by Yasunori Kano and colleagues, found that two of the living genera diverged recently on the mainstream clock, "despite their significant morphological differences," and proposed that the living group "might have radiated fairly recently from a cryptic lineage with ordinary rates of morphological evolution."⁴⁷  On that reading the living monoplacophorans are not a frozen Devonian lineage but a recent branch of a lineage that left no fossils for most of the record.

Verdict.  Class-level conservatism of shell and muscle scars, conceded by the mainstream;  no fossil of the living family before the Pleistocene;  and molecular evidence that the living genera are a recent radiation.

Ginkgo

The claim.  A Creation magazine article says that "Ginkgo biloba trees have manifested no detectable change in as much as 60 times that timeframe!", comparing the tree with the five to seven million years usually given for the human and chimpanzee lineages.⁴⁸

By rank.  This is the case where the young-earth argument reaches furthest down.  Dana Royer, Leo Hickey and Scott Wing write that "Undoubted remains of the genus Ginkgo" first appear in the Early Jurassic, "making it the oldest extant genus among seed plants."⁸  They call the fossil leaf species Ginkgo adiantoides "indistinguishable from modern G. biloba,"⁸ note that it "first appeared in the Early Cretaceous and was relatively common during the Late Cretaceous and Paleogene,"⁸ and conclude from their study of 51 fossil sites that "This marked stability in ecology supports previous work, based upon morphology alone, that G. adiantoides and G. biloba are conspecific."⁸  The species identity is inherited from earlier comparisons of leaves and cuticles;  Royer and colleagues did not measure it themselves.

Anatomy.  The genus did change.  Zhiyan Zhou and Shaolin Zheng described an Early Cretaceous Ginkgo from the Yixian Formation of China that filled a gap "during which the modern ovulate organs evolved from the Jurassic type."  Its seed-bearing organs "were more like those of the present-day Ginkgo biloba than those of the primitive Jurassic type, indicating that their morphology has changed little for over 100 Myr."⁴⁹  So the Jurassic members of the genus had different reproductive organs, and the stasis runs from the Early Cretaceous on.  The "200 million years" of unchanged morphology quoted in two ginkgo genome papers is not supported by the abstract of the paper one of them cites for it.⁵⁰

Genome.  A study that resequenced 545 ginkgo trees calls the tree "an enigmatic living fossil" and treats its stasis as "A mystery that has intrigued scientists," without measuring a rate.⁵⁰

Verdict.  The strongest case on the list.  The genus runs from the Early Jurassic with changed reproductive organs;  the leaf species that specialists treat as the living species runs from the Early Cretaceous.  "No detectable change" is true of the leaves since the Early Cretaceous and false of the genus since the Jurassic.

The Wollemi pine

The claim.  Creationist sources disagree with each other about what was claimed.  Ken Ham wrote that the tree "was thought to have become extinct with the dinosaurs, but was then found living alongside people in this present world!"⁵¹  An Answers in Genesis page in 2023 said it "was claimed (on the basis of the fossil record) to have gone extinct 150 million years ago."⁵²  A 2026 Answers in Genesis page says instead that "the last known fossil specimen was dated at 91 MY."⁵³  The New South Wales parks service, the source of the last figure, says the species was "thought to have gone extinct around 2 million years ago."⁵⁴

By rank.  The tree was described in 1995 as "a new species in a new genus."⁵⁵  No leaf, cone or wood fossil is assigned to Wollemia in the Paleobiology Database.  The fossil case rests mainly on pollen of the form genus Dilwynites, "the fossil pollen type that closely resembles the pollen of modern Wollemia and some species of its Australasian sister genus, Agathis," whose closest form reaches back to the Turonian, early in the Late Cretaceous.⁵⁶  The pollen specialists add that the nearest living relatives of the Dilwynites plants "are best regarded as both Wollemia and Agathis."⁵⁶  A 2023 genome preprint calls the tree "a southern hemisphere conifer morphologically unchanged since the Cretaceous," and in the same introduction concedes that "it is unclear to what degree the Dilwynites group also comprises the morphologically close Agathis based on palynological evidence."⁵⁷

Verdict.  An araucarian pollen type that includes Wollemia-like grains reaches the Late Cretaceous.  The genus and species have no securely identified fossil.  "Unchanged since the Cretaceous" outruns the evidence the preprint itself cites.

Others on the list

1. Dawn redwood. Metasequoia fossils run in the database from the Cenomanian, early in the Late Cretaceous, to the Pliocene, and the genus was named from fossils in 1941 before the living tree, which "was discovered during the 1940s" in China, was matched to it.²⁶⁵⁸ Royer and colleagues note that the fossil M. occidentalis "has remained largely unchanged morphologically throughout the Cenozoic" and "is represented today by M. glyptostroboides," a different species.⁸ Genus-level persistence, species-level change.
2. Laotian rock rat. Laonastes was described in 2005 and placed in the family Diatomyidae, "known from early Oligocene to late Miocene sites" in South and East Asia, "after a temporal gap of roughly 11 million years" on the conventional timescale.⁵⁹ Family-level persistence; no fossil of the living genus.
3. Tadpole shrimps. Triops has "an extremely conserved, yet complex, bauplan with extant species indistinguishable from fossils of Triops from the Triassic."⁶⁰ The same study found from DNA that the living species diversified recently and concluded that tadpole shrimps "can be regarded as 'living fossils' only on the grounds of morphological conservatism, not on the basis of limited diversification or relict status."⁶⁰

The cases side by side

Part 3: Schopf's objection, that soft-part change is invisible

The argument in his words

At the 1980 Field Museum conference on macroevolution, the meeting from which creationists still quote Niles Eldredge, the University of Chicago paleontologist Thomas Schopf argued the other way.  The New York Times wire report reads:  "He contended that species may not be as static as they seem.  Fossils, he noted, represent only the hard parts of the organism, such as bones or shells.  Schopf argued that the soft parts, which are not fossilized, may be undergoing significant gradual changes that the paleontologist can never see.  He recalled an instance in which the shells of certain marine organisms looked identical and were classified in the same species until live specimens were found.  Their internal organs differed so much that they had to be regrouped as separate species."¹  Then:  "'The so-called living fossils that people like to cite as examples of evolutionary stasis are largely a myth,' Schopf said.  For example, he noted, the species of horseshoe crab common today is unknown in the fossil record."¹

Schopf had published the core of the idea five years earlier with David Raup, Stephen Jay Gould and Daniel Simberloff.  Their paper argued that "more complex forms appear to change more rapidly," because a paleontologist can only split species and genera on characters he can see.  In a computer simulation, "Groups based on a few characters are longer lived and are commonly polyphyletic in comparison with groups based on many characters."⁶¹  A simple shell, on this view, offers too few characters to register change, so a lineage of simple shells will look ancient whatever it is doing.

Schopf's target was punctuated equilibrium, which used stasis as evidence that species change little between splitting events.  He was defending gradual change.  The companion article Trends in the Fossil Record covers that debate;  here the question is what his argument does to the living-fossil claim.

What it gets right

1. At species rank he was right about his own example, and it has held up. The living horseshoe crab species still has no fossil record, and the only extinct species still accepted in Limulus is a Late Cretaceous one (Part 2).
2. Cryptic species are real in exactly these groups. Genome-wide markers split the living nautiluses into species that do not match the shell-based names.⁴⁴ Tadpole shrimps that look alike turned out to be recent and diverse.⁶⁰ Two living monoplacophoran genera that diverged recently on the molecular clock show "significant morphological differences" from each other.⁴⁷ Shell identity does not guarantee species identity, which was Schopf's point.
3. His complexity argument predicts the database's long "genera." The Paleobiology Database still files more than a thousand tongue-shaped shells from the Cambrian to the Pleistocene as "Lingula," and its "Nautilus" records start in the Pennsylvanian, though the oldest species-level record a specialist has confirmed for the living genus is Late Eocene.²⁶⁴² When a specialist revised the lingulids, the Paleozoic and Mesozoic "Lingula" became other genera.⁷ Simple shells produced long-lived, mixed genera, as the 1975 paper predicted.

What it misses

1. It does not reach the rank the creationists argue at. Schopf showed that "unchanged species" is usually a myth. The young-earth claim is that the kind, roughly the family, has not changed (Part 1). A horseshoe crab whose living species is a "modern evolutionary development," in Schopf's phrase, is still a horseshoe crab, and on a young-earth reading a few thousand years of change within the family is expected.
2. Invisible change is also invisible stasis. Where soft parts cannot be seen, the fossil record cannot show that they changed either. Schopf's argument makes the soft-part question unanswerable from most fossils; it does not answer it.
3. The rare internal organs that did fossilize point both ways. One specimen of the Carboniferous horseshoe crab Euproops danae, from Mazon Creek in Illinois, preserves its brain and nerve cord. Bicknell and colleagues found that "the general prosomal synganglion organization has remained essentially unchanged in horseshoe crabs for >300 m.y., despite substantial morphological and ecological diversification in that time."¹¹ The coelacanth's lung went the other way. Mesozoic coelacanths such as Axelrodichthys carried a large lung wrapped in bony plates, which probably helped regulate its volume; in the living Latimeria the lung is "clearly vestigial" in adults and its plates are "a rudimentary anatomical structure."¹² Both results sit at family rank or above. Euproops belongs to an extinct family, and the same paper reads the nervous-system layout as shared by all chelicerates back to the Cambrian,¹¹ so it is a feature of the whole body plan, not of horseshoe crabs alone. The two organs test Schopf's general claim that internal anatomy can change unseen, and they show one organ kept and one reduced. Neither reaches his species-level point.
4. The genomes split as well. Gar genomes evolve far more slowly than any other vertebrate's,³² while the Lingula genome turns over its gene families faster than any bilaterian compared with it¹⁰ and horseshoe crabs duplicated their whole genomes.³⁹ The tuatara's ancient DNA was read as showing that "rates of neutral molecular and phenotypic evolution are decoupled,"³⁷ and Casane and Laurenti stress that molecular and morphological rates "are under different constraints and they are not necessarily correlated."³¹ So a slow genome does not prove slow soft anatomy, and a fast genome does not prove fast soft anatomy.
5. Hard-part conservatism has since been measured, not assumed. The coelacanth, tuatara and horseshoe crab studies in Part 2 counted characters and measured shapes across dozens of species. Their conservatism is a finding about the parts that fossilize, and it stands whatever the soft parts did.

Verdict on Schopf.  He is right that a living fossil is almost never the same species as its fossils, and his 1975 complexity argument explains the database's longest "genera."  He does not touch the creationist claim, which is pitched at family rank.  The genome data that might have tested his soft-part hypothesis are too decoupled from anatomy to decide it, and the two internal organs that have fossilized, a horseshoe crab nervous system and the coelacanth lung, show one kept and one reduced.

Part 4: What long stasis predicts on each model

On the mainstream model

Evolution, as the mainstream states it, never predicted that every lineage keeps changing.  Darwin's own explanation of living fossils was ecological:  they survived "from having inhabited a confined area" with "less severe competition."¹⁹  The modern explanations of stasis, chiefly stabilizing selection that holds a trait near an optimum, are set out and weighed in Trends in the Fossil Record, Part 1, which concludes that stasis is real and weaker than young-earth writers claim, and that the bound on change has no accepted explanation.  Vaux and colleagues put the mainstream reading of a living fossil plainly:  "limited morphological change over time might not indicate evolutionary conservatism, but trait fluctuation around a mean state due to constraining selection."³⁶  A simulation study by Dominic Bennett and colleagues adds a lineage-level condition:  trees shaped like the real tree of life arise only when "the most evolutionarily distinct tips have lower rates of speciation and extinction."⁶²

The cost is the length of time.  Uyeda and colleagues state the puzzle from the mainstream side:  field studies routinely see "substantial phenotypic change on a timescale of a few to a few dozen generations," yet "paleontologists working on much longer timescales have recognized an overwhelming prevalence of evolutionary stasis."⁶³  Philip Gingerich's compilation of measured rates gives the scale:  "about one-tenth of a standard deviation per generation," which he calls "more or less an upper bound."⁶⁴

That rate can be applied to the coelacanth for scale.  Latimeria waits at least 54 years between generations.⁶⁵  On the conventional timescale, the 66 million years since the last Cretaceous coelacanth are about 1.2 million generations.  A body measurement that wandered at random by Gingerich's upper rate each generation would drift about 110 standard deviations in that time.  On a young-earth timescale of about 4,500 years since the Flood, the same measurement gets about 83 generations and drifts about one standard deviation.  Whatever the true step size, the long clock gives about 120 times as much wandering as the short one.¹³  So on the mainstream model, a coelacanth that still looks like its Cretaceous relatives must have been held in place, generation after generation, for over a million generations.  Stabilizing selection is the proposed holding force.  It is well documented over short spans.  Why it holds some lineages for tens of millions of years and releases others is not known, and the trends article records that the mainstream literature itself calls the bound unexplained.

The illustration uses an upper bound, so a skeptic may fairly ask what mutation and drift alone would do.  Michael Lynch, a population geneticist, made that comparison for mammal skulls in 1990.  He found that "The rate of morphological divergence of almost all lineages, including the great apes, is substantially below the minimum neutral expectation," and concluded that "stabilizing selection is a predominant evolutionary force keeping the long-term diversification of lineages well below its potential."¹⁴  He also found that lineages diverge at about the neutral rate just after they split, "and that this rate declines over evolutionary time,"¹⁴ the same falling rate over longer intervals that Gingerich's compilation shows.  So the conclusion does not hang on the illustration.  On the mainstream's own measurements, lineages left to mutation and drift would change more than real ones do, and something holds them back.  The mainstream's answer, stabilizing selection, is a real and measured process;  what is still missing is an account of why it holds a coelacanth's skeleton for over a million generations.

The mainstream also does not predict which lineages will stall.  Lidgard and Love warn that "Assuming that sustained extinctions and slow evolutionary rates are uniformly characteristic of living fossil groups is also problematic."²¹  The ginkgo resequencing study calls the tree's stasis "A mystery that has intrigued scientists."⁵⁰  Each case is explained after it is found.

On the young-earth model

Young-earth writers state stasis as a prediction, but a hedged one.  Answers in Genesis says that "Stasis is predicted by creationists as one of the possible outcomes of variation within the created kinds as time progresses."⁵³  A 1997 Creation article sets out the whole prediction for fossils:  "Some fossil creatures would be somewhat different from their living counterparts due to genetic variation within the kind since the Flood, but still basically the same kind";  some "would not be represented among living things today";  and "Some fossils would be exactly the same as present-day kinds, just as shown in this example, because in reality, only a few thousand years separates the two specimens shown, not millions of years."¹⁵

The timescale is the young-earth model's real advantage.  On that model every fossil coelacanth, ginkgo and horseshoe crab in Flood rock lived within a few thousand years of its living relatives.  Long stasis then needs no holding force.  The arithmetic above shows how much less room for drift the short clock allows.

But the same model needs fast change, and its own writers say so.  Jeanson and Lisle place the created kind near the family, so "speciation has occurred post-Creation and post-Flood."  The rocks they assign to the post-Flood period hold so much new diversity that "a massive burst of speciation took place in just a few hundred years," and they call explaining it "the most challenging explanatory task that the YEC species' origins model faces."²⁴  John Whitmore reads the living-fossil counts the same way:  "organisms changed quickly after the Flood (producing many new genera within Genesis kinds), and once they became well-adapted to one of the many new niches after the Flood, change stopped."¹⁸  So the young-earth model predicts both fast diversification within kinds and stasis, and as stated it does not say which kinds do which.  The 1997 three-outcome list covers every possible fossil, which means no fossil can count against it.

The short clock also has to allow the differences that do exist.  Latimeria differs in body shape from its closest Cretaceous relatives (Part 2).  In about 83 generations, a directional change at Gingerich's upper rate could move a trait about eight standard deviations,¹³ so on the young-earth clock that difference is affordable too.  Neither the stasis nor the change strains the short clock.  What strains it is the volume of new genera after the Flood, which is a separate problem that Jeanson and Lisle name.

What would tell the models apart

Stasis itself does not.  Both models can absorb a lineage that stalls and a lineage that changes.  One pattern does bear on the question.  In two of the best-measured cases the living form resembles its younger fossils more than its older ones.  Latimeria sits close to its Mesozoic relatives and far from the Devonian forms;  living ginkgos match Cretaceous seed organs and not Jurassic ones.  On the mainstream model that is what descent with slowing change predicts.  The horseshoe crabs do not fit the pattern.  Bicknell's measurements put the living shape close to both the Jurassic forms and the early Paleozoic ones, with the unusual shapes in between, and the unusual shapes go with forays into marginal habitats rather than with position in the rocks.⁴¹⁴¹  Its authors explain this as repeated convergence on one marine form.  That takes the horseshoe crabs out of the descent-with-slowing-change argument.  It does not by itself favour a Flood model either, which would need the marginal-habitat forms buried at two separate levels, Carboniferous and Triassic, with marine forms below, between and above them.  No young-earth source read for this article has taken up the horseshoe crab case.  On the young-earth models now in use, which end the Flood at the top of the Cretaceous or higher, all these fossils are near-contemporaries buried in sequence during one year, so the pattern has to come from the order in which communities were buried.  That is the ecological-zonation argument weighed in The Order of the Fossil Record, Part 5.  No young-earth source read for this article applies zonation to living fossils in that way.

An older young-earth model does not need zonation here.  In 1990 Joachim Scheven divided the fossil record into "two pre-Flood and two post-Flood chief ecosystems" and put the whole Mesozoic after the Flood, "an episode which modern man pleases to call the 'Mesozoic'."⁶⁶  He called it "a time of considerable genetic drift" in small populations, and wrote:  "Here, at last, we have true evolution, i.e. the unwrapping of genetic potential that lay hidden in many created species."⁶⁶  On that scheme the Devonian coelacanths died in the Flood, the Mesozoic coelacanths and ginkgos descend from its survivors, and the living forms should look most like the youngest fossils, as they do.  Scheven did not apply his scheme to coelacanths or ginkgo seed organs;  the application is this article's.  The scheme's cost is its Flood boundary.  It leaves the dinosaur beds, the Chalk and most of the sedimentary record to the centuries after the Flood, and the four authors who revived a Palaeozoic boundary in 1996 withdrew it in 2000 (The Flood/Post-Flood Boundary, Part 1).  So the pattern fits a young-earth history without zonation only on a boundary that no active young-earth camp now holds.

Part 5: The gaps, and why they bind both sides

What has already been settled

A Lazarus taxon is a group that disappears from the rock record and reappears higher up or alive today.  The Order of the Fossil Record, Part 3, shows that such gaps only ever extend a group's range upward and never put it below something it must postdate, so they do not disturb the order of the fossil record.  The same part makes the point that "absence of evidence" is weak in both directions:  a side that uses Lazarus taxa to show the record is patchy cannot also use gaps between fossil forms to show the record is complete.  This part does not repeat those arguments.  It asks a narrower question that has not been asked before:  is each living-fossil gap large enough that it cannot be put down to ordinary bad luck in collecting?

How the young-earth side uses the gaps

Answers in Genesis makes the coelacanth an object lesson in method:  "For decades, evolutionists said that coelacanths went extinct 65 million years ago, because above that level in the geological record there were no coelacanth fossils.  The absence of evidence was considered evidence of absence, until 1938 when they found the fish swimming off the coast of Madagascar."⁵²  (The 1938 fish was found at East London, South Africa;⁹ Madagascar is an error the page shares with Ken Ham's article and an ICR page.)  Michael Oard turns the gap against the long timescale:  "Living fossils challenge evolution and uniformitarianism because there are often vast time gaps between their last appearances and the present."⁶⁷  Werner applies the same principle to his own side:  horses, humans, bats and goats are missing from dinosaur rock, "The same is true for the theory of creation," and their presence then is "implied by the large number of other modern animals and plants that have been found."²

The methodological point is correct.  Absence from a rock layer does not prove absence from the world at that time.  What it does not do is tell either side how surprising a particular absence is.  That can be computed.

The gaps, priced

The test asks how many fossil sites inside a gap should contain the living fossil if it had stayed as common, relative to its neighbours, as it was just before the gap.  For coelacanths the neighbours are ray-finned fishes;  for tuatara relatives, lizards and snakes;  for monoplacophorans, snails;  for horseshoe crabs, crabs, shrimps and lobsters.¹⁶

These are not sampling accidents.  If coelacanths had stayed as common as they were in Late Cretaceous seas, the chance of finding none in 1,686 later marine fish sites is about one in 10²⁴.  For monoplacophorans it is smaller still.  Only the horseshoe crab gap, with about seven expected, could plausibly be bad luck, at odds of roughly one in 700.  Each lineage either became rare or retreated into places the rock record seldom samples.

The coelacanth case shows which.  The Cretaceous coelacanths were not deep-sea fish.  Of the 61 Cretaceous coelacanth sites in the database, the great majority are river, lake, estuary, lagoon, coastal and shallow-sea deposits;  one is offshore and none is from the continental slope or a deep basin.¹⁶  Later rocks hold hundreds of river, lake and coastal sites full of ray-finned fishes, and no coelacanth.  The two living species of Latimeria "live in the deep ocean,"³¹ the one setting the later record barely samples:  there are only 63 deep-water fish sites in the whole Cenozoic in the database.  So the coelacanths that survived are, on either model, the ones that lived where fossils seldom form, and the ones that lived where fossils do form were gone.

What each side has to say

On the mainstream model, the survivors spent tens of millions of years in refuges:  deep slopes for the coelacanth, the abyssal plain for monoplacophorans, the islands of New Zealand for the tuatara, a rainforest canyon of the Greater Blue Mountains for the Wollemi pine.⁵⁴  That is consistent, but it is an explanation found after the fact, case by case.  Camila Cupello and colleagues state the coelacanth version in their own words:  a move into deep water "may help explain the survival of some coelacanths throughout the Cretaceous/Palaeogene (K/Pg) environmental crisis, whereas Late Cretaceous coelacanths inhabiting shallow waters disappeared," so that the gap "is merely due to a lack of Cenozoic fossil coelacanths so far, and scarce preservation of Cenozoic deep-water sediments."¹²  The table supports the first half and not the word "merely."  The shallow-water coelacanths did disappear, and it is their disappearance, not missed collecting, that leaves the later lake, river and coastal rocks empty.

On the young-earth model, the answer depends on where the Flood ends, which young-earth geologists dispute.⁶⁸  If the Flood ended at the top of the Cretaceous, the coelacanths survived it and then left no trace in any post-Flood lake, river or shore deposit for four and a half thousand years, while ray-finned fishes filled those deposits.  If the Flood ended near the top of the Neogene, as the Institute for Creation Research holds, the Cenozoic rocks are the Flood's receding phase.  ICR's Brian Thomas offers that model's explanation:  the Cretaceous "were about the high-water mark of the Flood, reached on Day 150," the layers above "buried mostly higher elevation land plants and creatures," and so "No wonder deep sea creatures are hard to find among the Earth's uppermost fossil layers."⁶⁹  The table tests that explanation.  It has something to it, because the Cenozoic does sample deep water poorly.  But it does not fit the fossil coelacanths, which lived in rivers, lakes and coastal seas, the very settings the receding-phase rocks preserve in abundance.

Both sides end with the same explanation:  the surviving lineages lived where rocks were not forming or are not exposed.  That is a reasonable explanation on either clock.  It is not evidence for either clock, because the gap is equally unexplained by sampling on both.

Why the gaps matter to both sides

1. For the mainstream, a priced gap is a warning against using incompleteness as a blanket answer. The coelacanth's absence is not a gap in collecting; the habitats it once lived in were sampled hundreds of times. An appeal to "the record is incomplete" for a missing transitional form carries weight only when the habitat in question can be shown to be poorly sampled.
2. For the young-earth side, a Lazarus gap is not evidence of a young earth. The coelacanth was absent from the later rock on the long clock and on the short one. The argument from silence that Answers in Genesis and Werner make is sound as a caution, and it cuts both ways: if absence proves little, then the absence of human fossils from dinosaur rock, and of transitional forms between kinds, also proves little.
3. Most living genera have no old fossils at all. The census in Part 1 shows that only 28% of living mammal genera, and fewer than 10% of living bird, reptile and amphibian genera, have any fossil older than the Pleistocene. A living group missing from older rock is the normal case. What makes the famous Lazarus taxa remarkable is not that they vanished but that they were once common and then vanished from habitats that stayed well sampled.
4. The gaps are shorter than the popular figures. The monoplacophoran gap is usually given as the whole span since the Devonian. The class's fossil record in the database runs into the Permian and possibly the Late Triassic, and a living species has been found in Pleistocene clays in Italy.⁴⁶ The tuatara line has Paleocene and Miocene fossils inside its supposed gap. Gaps shrink as collecting continues; the creationist writer Michael Oard, in the paper quoted above, makes the same observation about fossil ranges in general.

Objections and responses

1. "Paleontologists give fossils different names because they expect change. The differences are an artefact of evolutionary bias." Don Batten argues this, reporting Werner's view that "paleontologists tend to give the fossil forms quite different names to the living ones."⁷⁰ The naming runs both ways in the sources read here. Where specialists find no difference they lump fossil and living forms together: Royer and colleagues accept Ginkgo adiantoides as the living species,⁸ Taviani and colleagues put Pleistocene shells in the living Micropilina minuta,⁴⁶ and a 1996 restudy put an Eocene shell into Nautilus.⁴² Where they find differences they split, as Lamsdell did when he moved the Jurassic "Limulus" out of the living genus.⁶ The largest naming errors on record run the other way, toward too-long ranges: more than a thousand Paleozoic and Mesozoic shells were filed as "Lingula" until a specialist revised them. Werner himself concedes that even the dinosaur-era dragonfly, gar, coelacanth and horseshoe crab "were so different from the living forms" that they were given different "genus and species names indicating they could not reproduce with modern forms (if they lived at the same time)."²
2. "Evolution is about change, so stasis falsifies it." Darwin explained living fossils by survival in confined, less competitive habitats, not by a law of constant change.¹⁹ The objection does land on one point, which this article concedes: holding a lineage nearly still for over a million generations needs an active constraint, and the mainstream has no accepted account of why the constraint holds some lineages and not others (Part 4).
3. "The young-earth model predicted stasis, and stasis is what we see." The young-earth literature predicts both stasis and fast diversification within kinds, and its fullest statement of the prediction lists every possible outcome (Part 4). A prediction that no fossil could contradict does not gain support when a fossil fits it.
4. "The gap test assumes the coelacanth's habitat never changed. That is a straw man." It assumes exactly that, on purpose. The test measures how far the record departs from constant habitat, and the answer is that it departs by a factor no sampling accident can produce. The conclusion drawn is that the habitat did change, on either model.
5. "The Paleobiology Database is incomplete and biased toward North America and Europe, so the counts mean little." The bias is real and has been measured in the companion trends article: the United States alone holds 22% of the database's collections. But the gap tests compare a taxon with its neighbours in the same database, so the bias affects both sides of each ratio. The genus census gives lower bounds, because the database does not hold every published fossil. A factor-of-two error in any row changes none of the conclusions.
6. "Whitmore found 2,594 living-fossil genera in the Cenozoic, four times the Mesozoic count. That is the post-Flood burst." Whitmore counts any genus with a fossil older than 2.6 million years on the conventional scale and living members today.¹⁸ More Cenozoic genera pass that rule partly because Cenozoic rocks are younger, more widely exposed and more heavily collected, and his count is not corrected for collecting effort. His own footnote adds that "It is not yet possible to evaluate stasis at the species level" in the database.¹⁸ The count is consistent with a post-Flood burst and with ordinary turnover; it does not test between them.
7. "Ginkgo proves a living species unchanged since the Cretaceous." It is the best case, and this article says so. But the species identity rests on leaves and cuticles compared in earlier work, which Royer and colleagues cite rather than measure,⁸ and Schopf's warning applies to plants as well: the organs that fossilize best are not the whole organism. The genus did change its seed-bearing organs after the Jurassic.⁴⁹
8. "Molecular-clock dates, such as the recent divergence of living monoplacophorans, assume evolution and deep time." They do. This article reports them only as the mainstream estimate. The observation that does not depend on a clock is the one Kano and colleagues report alongside it: living monoplacophoran genera show "significant morphological differences" from each other.⁴⁷
9. "Absence of evidence is not evidence of absence (Ham, Answers in Genesis)." As a caution, that is right. The pricing in Part 5 shows how to turn the caution into a number, and the number for the coelacanth is large on both timescales. The same caution also weakens arguments from the absence of transitional forms, which is why The Order of the Fossil Record treats the two together.

Issues and unanswered questions

1. Schopf's 1984 review was not read. "Rates of evolution and the notion of 'living fossils'" (Annual Review of Earth and Planetary Sciences 12:245–292) sits behind a bot challenge. His six definitions are quoted here from Eldredge's chapter in the 1984 casebook, and his published argument from the 1975 abstract. The full review may answer the rank objection made in Part 3.
2. Several primary papers were read only in abstract. Kano and colleagues 2012 on monoplacophorans, Lemche's 1957 description of Neopilina, Lamsdell's 2016 horseshoe crab phylogeny, Hay and colleagues 2008 on tuatara DNA, Dawson and colleagues 2006 on Laonastes, Combosch and colleagues 2017 on nautiluses, Wani and colleagues 2008, Taviani and colleagues 1990, Saunders and colleagues 1996, and Zhou and Zheng 2003 were read as abstracts. So was Lynch 1990 on neutral rates of mammal skull evolution, which is quoted from its abstract only. Zhou's 2009 review, which two genome papers cite for "200 million years" of ginkgo stasis, and Crane's 2013 book were not obtained.
3. The tuatara anatomy is contested and the last word was not read. Herrera-Flores and colleagues replied to Vaux and colleagues in 2019; the reply returned an access error.
4. The coelacanth evidence conflicts within the mainstream. Clement and colleagues report a Devonian burst and later slowing of disparity; Christophe Ferrante and Lionel Cavin write that "the morphological disparity of this clade has remained remarkably stable" since it split from other bony fishes.⁷¹ Amemiya and colleagues find slow protein evolution; Brownstein and colleagues find ordinary rates; a 2026 gar genome paper, with Brownstein as a co-author, says gar substitution rates are "surpassed only by the coelacanth," the reverse of the 2024 ranking.⁷² The article reports these as open.
5. The gap pricing rests on database environment codes and on neighbour groups standing in for preservation chances. A deep-water coelacanth would be expected in only about five Cenozoic sites even at its old frequency. A finer test would compare deep-water fish faunas directly.
6. The young-earth side has published no test of its own gap mechanism. ICR's deep-water explanation makes a prediction that can be run against the marine record with sampling corrected; Part 5 runs a first version. No young-earth source read here defines the kind for any living-fossil group, or applies ecological zonation to the pattern in which living fossils resemble their younger fossils more than their older ones. Scheven's 1990 scheme, which makes the Mesozoic post-Flood, would explain the pattern without zonation, but its Flood boundary has been abandoned (Part 4).
7. A sibling article needs a correction. The Lazarus table in The Order of the Fossil Record gives the monoplacophoran record as ending in the Devonian, with a gap of about 380 million years. The database's monoplacophoran class (Tergomya) runs into the Permian and possibly the Late Triassic, and a living species is known from the Pleistocene of Italy.

Verdict

At family rank and above, living fossils are real, and the young-earth observation is fair.  Coelacanths, horseshoe crabs, nautiloids, lingulids, monoplacophorans and ginkgos keep their body plans through most of the record, and mainstream specialists concede it in print.  At genus rank most famous cases fail, and at species rank only the ginkgo's leaves cross into Mesozoic rock.  "Unchanged for hundreds of millions of years" is almost always a slide from a living species to a fossil order.

Long stasis fits a short timescale more cheaply than a long one.  Over about 80 generations a lineage needs no holding force to stay put;  over a million generations it does, and the mainstream's holding force, stabilizing selection, works without an accepted account of why it holds some lineages and not others.  But stasis does not decide between the models, because the young-earth literature also predicts fast change within kinds, and its fullest prediction covers every outcome.  The one pattern that bears on the question, living coelacanths and ginkgos resembling their younger fossils more than their older ones, fits descent directly and fits the Flood model only through an ecological sorting that no young-earth source has worked out for these groups.  Horseshoe crabs break the pattern:  their shape follows habitat rather than age, which neither model predicted.

The gaps are real rarity, not missing collections, on both clocks.  The coelacanth should appear in about 54 later marine fish sites and in about a dozen later lake and river sites, and it appears in none.  Both models must put the survivors in habitats the rocks did not sample, and neither gains from the gap.  Schopf was right that the living species are not the fossil species, and right that simple shells make long false genera;  he did not touch the creationist claim, which is made at the rank of the kind.

Sources