Trends in the Fossil Record

Gould, Stephen Jay.  “Evolution as Fact and Theory.”  Hen's Teeth and Horse's Toes, 253–262.  New York: Norton.  1981 (edition of 1983).  An essay first published in Discover in May 1981 and reprinted in 1983.  It answers the creationist arguments of its day one at a time, and argues that evolution is both a fact and a theory, that preserved transitions are uncommon but not absent, and that creationism cannot be falsified.  The article uses it for Gould's complaint about being quoted by creationists and for the sentence that follows it. ↩

1. 1a.

   Page 260:  The sentence before it supplies the meaning.  Punctuated equilibrium was proposed to explain trends, and a trend.  “is more like climbing a flight of stairs”.  than rolling up an inclined plane. ↩ ↩ ↩

2. 1b.

   Page 260:  The sentence immediately following it says transitional forms are generally lacking at the species level but abundant between larger groups.  Gould then names Duane Gish and Luther Sunderland personally as the creationists he is answering, and quotes both.  A second, earlier witness, John R. Cole, "Misquoted Scientists Respond," Creation/Evolution VI (Fall 1981), reproduced by the National Center for Science Education at https://ncse.ngo/misquoted-scientists-respond, gives a slightly different continuation: the same two sentences followed by "The evolution from reptiles to mammals … is well documented."  The Discover original carried that sentence, and the 1983 book reprint dropped it.  Both are Gould, and neither changes the meaning. ↩ ↩ ↩

Berean Archive.  What the paleontologists actually said.  n.d.  The full quotation audit for this series, with every verdict and every failure, covering 236 passages across eleven separate audits.  The article points readers to it as a companion page that includes the failures. ↩ ↩ ↩ ↩

Paleobiology Database.  Paleobiology Database.  n.d.  The database of published fossil occurrences, collections and taxa, counted here in full: 2,001,053 occurrences and 277,490 collections.  The country-code split is not cosmetic: a query that never needs the collections table silently drops 27,232 whole collections, 225,113 records or 11.25% of the database, so the unsplit query returns 1,775,940 occurrences instead of the full count.  The GBIF mirror of the database is stale, reporting 1,310,925 occurrences, so the direct count should be cited.  Every analysis uses one rule, because each occurrence carries an age bracket, not a point estimate: a claim that something is older than a boundary is tested against the bracket's minimum age, and a claim that it is younger against its maximum; a collection recorded only as "Paleozoic" spans 287 million years.  Measured age resolution: 50.6% of collections pin to a bracket of 5 million years or less, 83.3% to 10 million years or less, 70.1% to a single ICS stage, and 93.4% of occurrences (1,868,154) carry a bracket of 15 million years or less.  The article uses it because it computes its numbers from the database itself rather than quoting someone's summary. ↩ ↩

1. 3a.

   Twenty-four boundary tests, each asking whether any occurrence's entire age bracket falls past the boundary.  Below-first-appearance counts: mammals in Paleozoic rock 0; birds in Triassic or older rock 0; dinosaurs in Paleozoic rock 0; land vertebrates in Silurian or older rock 0; vertebrates and chordates in Precambrian rock 0 each; primates in Early Cretaceous or older rock 0; whales in Mesozoic rock 0; great apes in Eocene or older rock 0; grasses in Jurassic or older rock 0; trilobites in Precambrian rock 0; ammonoids in Ordovician or older rock 0; amniotes in Devonian or older rock 1; insects in Ordovician or older rock 1.  The amniote hit is occurrence 2050687, identified in the database itself as "Dipnoa indet.", a lungfish, in Early Frasnian rock.  The insect hit is occurrence 1661280, Aulocopsis simplex, from the Ordovician Volkhov Formation; Aulocopsis is a sponge genus name bound to an insect homonym.  Above-last-appearance counts total 19, of which 3 are the Mesophyllum coral/alga homonym, 1 is a nomenclatural artifact, 1 shows the test boundary itself was wrong (dendroid graptolites are known into the Carboniferous), and 14 are the two best-known reworking cases in the literature: the Gulf Coast Danian ammonites of the Clayton and Kincaid formations, every one from a collection whose age bracket begins at the boundary itself, and Fort Union theropod material from the Hell Creek transition.  The one genuinely open set is 25 pre-Jurassic angiosperm records, all leaf, seed or dispersed-pollen morphogenera whose affinity is disputed in the primary literature.  A further 85 apparent Paleozoic "flowering plants" are form taxa and vanish entirely under the database's regular-preservation filter, so anyone quoting a raw count without that filter is quoting a filing convention. ↩ ↩ ↩ ↩

Berean Archive.  The Order of the Fossil Record.  n.d.  A Berean Archive article that develops the order results summarised here, including the hydrodynamic and ecological sorting models offered to explain them.  The article uses it to separate the order, which is not in dispute, from its explanation, which is. ↩ ↩

Wise, Kurt P.  “Towards a Creationist Understanding of 'Transitional Forms'.”  CEN Technical Journal (now Journal of Creation) 9(2):216–222.  1995.  A 1995 paper in a creationist journal by Kurt P. Wise, a young-earth creationist paleontologist.  It argues that creationists should not spend scarce resources on the transitional-forms argument, proposes new terms for kinds of intermediate fossil, and goes through Darwin's expected intermediates one at a time, finding that three of the four kinds are confirmed.  The article uses it for those concessions and for the whale series, which Wise calls the largest challenge to his own model.  His credential line is printed in the paper itself at p. 222. ↩

1. 5a.

   Page 219:  Wise states that creationists need to accept the fact of the confirmed stratomorphic intermediates, and that it certainly cannot be said that traditional creation theory expected or predicted any of these fossil finds.  It must not be quoted without its companion at p. 218, which rules that the first of Darwin's expectations fails.  His footnotes source every example to mainstream literature and not to creationist works: Stewart & Rothwell 1993 for Baragwanathia; Gould's Wonderful Life pp. 321–323 for Pikaia; Carroll's Vertebrate Paleontology and Evolution (1988) p. 467 for Purgatorius, p. 473 for Proconsul, pp. 527–530 for the phenacodontids and pp. 527–549 for the Cenozoic mammal series; Hopson 1994 for the mammal-like reptiles; Ostrom 1994 for birds; Thomson 1994 and Ahlberg & Milner 1994 (Nature 368:507–514) for tetrapods; Gingerich 1994, Gould 1994 and Zimmer 1995 for whales.  A reader can check every one of Wise's claims without touching a creationist source. ↩ ↩

2. 5b.

   Page 218:  Under his own heading, "It is a Very Good Evolutionary Argument," Wise restates Darwin's expectations as four kinds of stratomorphic intermediate and rules on all four.  The first fails: he writes that the commonness of inter-specific stratomorphic intermediates.  “has been the most disappointing for classical Darwinists”.  , and that the current lack of any certain example is what drove the acceptance of punctuated equilibrium.  The other three succeed. ↩ ↩

3. 5c.

   Page 218, column 1:  “it is important that creationists NOT use 'transitional form' in their argumentation.”  On the same page he corrects a specific creationist misuse of Gould, writing that Gould's different definition of the term has led some creationists to conclude wrongly that Gould does not regard Archaeopteryx as a transitional form. ↩ ↩

4. 5d.

   Page 219, column 2:  “At this point in time, the largest challenge from the stratomorphic intermediate record appears to this author to come from the fossil record of the whales.”  “At present creation theory has no good explanation for the fossil record of whales.”  Both quotations are verbatim.  His reasons are given in detail: the archaeocete sequence is both stratigraphic and morphological; a land-to-sea transition cannot be explained by his transgressive-Flood ordering argument; and whales appear only in sediments his model treats as post-Flood. ↩ ↩

5. 5e.

   Page 220, column 1, section item (d):  Wise suggests that sightings of Nessie and similar lake creatures.  “might be living examples of the archaeocete Basilosaurus”.  This is dead by his own movement's later ruling: CMI's list of arguments not to use now places "There are living dinosaurs and plesiosaurs today" among arguments that are doubtful and inadvisable, with the reasoning that enough years have passed and "photographs can now be taken with a smartphone."  Anyone citing Wise on whales should know that the Loch Ness paragraph sits on the facing page, because a hostile reader who opens the PDF will find it. ↩

Sarfati, Jonathan.  “Arguments we think creationists should NOT use.”  Creation Ministries International.  Feb 7, 2006.  The Creation Ministries International list of arguments creationists should not use, sorted into a "definitely not" tier and a "doubtful" tier, dated 7 February 2006 and updated since.  The history matters and is usually got wrong.  Answers in Genesis published the list first, and CMI, formerly Answers in Genesis-Australia, separated from AiG-US in 2005–2006 and kept it.  The two texts share their preamble almost verbatim, but AiG no longer carries the list: its topic hub has nineteen articles, whose fossil-relevant content is two items, so "Answers in Genesis says creationists should not use X" is, for most fossil items, no longer true, while CMI says it.  The article uses it for the arguments the movement has retired and for the ones it has not. ↩ ↩

1. 6a.

   Section What arguments are doubtful, hence inadvisable to use?:  “Since there are candidates, even though they are highly dubious, it's better to avoid possible comebacks by saying instead: 'While Darwin predicted that the fossil record would show numerous transitional fossils, even 150 years later, all we have are a handful of disputable examples.'”  The 2005 Answers in Genesis version is identical except that it reads 140 years.  The Paluxy entry moved between tiers between the two versions, and in both directions at once: AiG placed it in the "definitely not" block, while CMI places it under "doubtful", but CMI's text is harder, inserting "Research to date has not been promising about the tracks, however."  On Archaeopteryx, CMI says: "Archaeopteryx was genuine (unlike Archaeoraptor, a 'Piltdown bird'), as shown by anatomical studies and close analysis of the fossil slab.  It was a true bird, not a 'missing link'."  On Baugh, CMI says: "we think that he's well meaning but that he unfortunately uses a lot of material that is not sound scientifically."  On Dubois and Java man, CMI says the claim misreads him, "as Stephen Jay Gould has shown"; note what that costs the movement rhetorically: the retraction is sourced to Gould. ↩ ↩

2. 6b.

   Not on either list: Precambrian pollen (the Roraima Formation and the Salt Range material), the Moab skeletons, the Meister "trilobite in a sandal print," and the London Hammer and the artifacts-in-coal family generally, except a narrow entry on gold chains in coal, which retires the evidential form of the claim without naming any specific artifact.  Also absent, and therefore still live within the movement: the horse series as a discredited icon, "the Cambrian explosion refutes evolution," polystrate fossils, "the geologic column exists nowhere in full," and coelacanth-as-living-fossil arguments.  Absence from the list is neither endorsement nor retirement; it means only that neither organisation has ruled. ↩

Morris, John D.  “The Paluxy River Mystery.”  Impact (151).  Institute for Creation Research.  Jan 1, 1986.  “Even though it would now be improper for creationists to continue to use the Paluxy data as evidence against evolution, in the light of these questions, there is still much that is not known about the tracks and continued research is in order.”  “coloration phenomena only, with no impressions, in most cases”.  “claims neutrality on the creation-evolution question”.  A short article, published 1 January 1986, in ICR's Impact series by John D. Morris, ICR's president, which withdraws the institute's earlier use of the Paluxy man-tracks as evidence against evolution.  Morris had written the 1980 book Tracking Those Incredible Dinosaurs, one of the two main sources of the claim.  He credits the discovery to Glen Kuban, who noted the staining in 1982 and invited the original film team back to the site in October 1985, and says he claims neutrality on the creation-evolution question.  Morris gives a specific, checkable mechanism: the elongated "Taylor Trail" impressions developed three-toed colouration patterns as a reddish stain spread, and the toes are "coloration phenomena only, with no impressions, in most cases."  The quotations below are verbatim.  One sentence weakens the retraction: the list of "mysterious points" immediately before the withdrawal includes the observation that a reddish stain can easily be applied to rock with readily available chemicals, an insinuation of tampering made without naming anyone and without evidence.  ICR also distances itself from Carl Baugh in the same piece.  The article uses it because ICR printed its own retraction of a claim it had promoted, and credited a neutral investigator with the discovery that killed it. ↩ ↩

Mayr, Ernst.  What Evolution Is.  Basic Books, New York.  2001.  Page 14:  “Instead, he or she finds gaps in just about every phyletic series,”.  “Indeed the fossil record is one of discontinuities, seemingly documenting jumps (saltations) from one type of organism to a different type.”  A book by Ernst Mayr that sets out what evolution is.  The article uses his statement that a paleontologist finds gaps in just about every phyletic series, meaning a sequence of fossils thought to form one line of descent, together with the answer he gives on the same page. ↩

1. 8a.

   Page 14:  The next sentence is Mayr's answer, and it runs for the rest of the page.  It gives Darwin's appeal to the incompleteness of the record, the subduction and metamorphism of fossil-bearing strata, the small fraction of strata exposed at the surface, and the low probability that any organism fossilises at all.  It then names Archaeopteryx as a fossil that fills a large gap and confirmed a prediction that anatomists had already made.  Mayr adds:  “A few fossil lineages are remarkably complete,”.  Page 15:  The therapsid-to-mammal series that Mayr names with this sentence is given as a cladogram across the whole of the page.  Quoting the gaps sentence without this answer is indefensible, because it would misrepresent the book.  The answer names precisely two examples, Archaeopteryx and the therapsid-to-mammal series, and this site has separate articles for both. ↩ ↩ ↩

Simpson, George Gaylord.  Why and How: Some Problems and Methods in Historical Biology.  Pergamon, Oxford.  1980.  Chapter 1:  “Gaps among known species are sporadic and often small”.  “Gaps among known orders, classes, and phyla are systematic and almost always large”.  A book in which Simpson sets out the problems and methods of historical biology, the study of how life has changed over time.  The article uses his statement that gaps among known species are sporadic and often small, while gaps among higher groups are systematic and almost always large, together with the explanation he gives for them. ↩

1. 9a.

   Chapter 1:  “There is little doubt that the rise of a higher category has frequently involved exceptionally fast evolution,”.  In such transitions, Simpson says, the populations involved were probably.  “notably smaller in size and much more restricted in distribution”.  His answer to his own observation is small, localised, fast-evolving populations.  Eldredge and Gould gave the same answer twelve years earlier.  Quoting his description of the gaps without this explanation would be quote-mining. ↩ ↩ ↩

Erwin, Douglas H.  “Evolutionary uniformitarianism.”  Developmental Biology 357(1):27–34.  2011.  A compilation of when marine invertebrate phyla and classes first appear in the Ediacaran and Cambrian rocks, with an argument that the clumped pattern of body forms partly records how variation was generated over time.  It is an evolutionary-developmental paper, written for a meeting on that subject at Ben-Gurion University, and its whole argument turns on developmental mechanism.  The article uses it for Erwin's statement that the pattern remains one of the open questions in evolutionary biology, and for his argument about sampling. ↩ ↩

1. 10a.

   Section Discussion:  The sentence about non-paleontologists explaining the pattern away as a bias is in the Discussion, six pages later, and not in the introduction.  The two sentences after it give Erwin's own alternative, that combining the morphological pattern with comparative studies of developmental gene-regulatory networks yields.  “a new, non-uniformitarian view of evolution”.  Splicing the two passages together with an ellipsis reverses the function of the second passage. ↩ ↩

2. 10b.

   Section Introduction:  The printed proof reads "one of most important", an uncorrected typo in the proof itself.  “The ubiquity of morphological discontinuities between clades of organisms has troubled evolutionary biologists since Cuvier and Darwin and remains one of most important questions in evolutionary biology.  Why is it that the distribution of morphologies is clumpy at virtually all scales? …  Gradations in form are even less common at higher levels of the Linnean taxonomic hierarchy.”  The same paragraph lists the three standard classes of explanation.  The first is.  “neo-Darwinian extrapolation in which morphological disparity reflects gradual divergence followed by the disappearance, through extinction or non-preservation, of intermediate forms.” ↩ ↩

3. 10c.

   Section Introduction:  After his summary of the quantitative disparity literature, Erwin writes:  “These studies quantitatively substantiated Gould's intuitive conclusion:  morphologies really are unevenly distributed at the origin of a clade.  In the majority of cases studied, morphologic disparity greatly exceeds taxonomic diversity in the early history of a clade.  This result could only be the result of sampling artifacts if one were to posit that unusual morphologies were more likely to be preserved and recovered than morphologies similar to one another, which is hard to credit.  Such studies have demonstrated that the apparent gaps between morphologies are not simply due to extinction of once-intermediate forms.” ↩

4. 10d.

   Section Abstract:  Erwin is an evolutionary biologist proposing an evolutionary mechanism, and nothing in this paper supports special creation.  “I argue that the distribution of morphologies may be, in part, a record of the time-asymmetric generation of variation.  Evolution has been implicitly viewed as a uniformitarian process where the rates may vary but the underlying processes, including the types of variation, are essentially invariant through time.  Recent studies demonstrate that this uniformitarian assumption is false.”  Section Discussion:  “the formation of the distinctive kernels associated with regional pattern formation in the developing embryo closely coincides with the origin of phyla and classes during the Ediacaran–Cambrian.  Indeed the formation of these kernels and their recursive wiring has been chiefly responsible for the formation and subsequent stability of these distinctive clades.” ↩ ↩

5. 10e.

   “For example, 21 different classes have been described for the Phylum Echinodermata, but many of these are short-lived, if morphologically distinctive, groups known only from the Cambrian or Ordovician periods.  These are recognized as classes only because they were described many decades ago.  Today many of these echinoderm clades would be recognized as unranked stem groups within the Echinodermata.  This difference in no way changes their morphologic distinctiveness … but reflects the evolution of taxonomic bookkeeping.”  He also states his own tabulation rule:  “Stem groups within phyla were tabulated as classes, and stem groups within classes are tabulated as orders.” ↩ ↩ ↩

6. 10f.

   Section Results:  Erwin's own compilation shows a major origination pulse in Cambrian Stage 1, then first appearances of many clades in Stage 3, then a smaller pulse at the Burgess Shale.  “The only later occurrence of a durably skeletonized phylum is the Bryozoa in Cambrian stage 9”.  “Twelve phyla are known only from the Recent and four phyla first occur in the fossil record after the late Cambrian.”  His explanation for the late arrivals is that.  “All of these groups are non-durably skeletonized and evidence from molecular phylogenies and molecular clocks suggests that the lineages they represent arose by the early Paleozoic, so it is not surprising that they have a poor fossil record.”  This is a preservation argument that Erwin makes against his own headline pattern. ↩

Holtz, Thomas R. Jr.  “GEOL 204 lecture links page.”  University of Maryland.  n.d.  “The punctuations occur at the level of species; directional trends (on the staircase model) are rife at the higher level of transitions within major groups,”.  A lecture links page for a University of Maryland geology course.  Its third epigraph runs the "infuriating to be quoted" sentence straight into a second sentence, inside one pair of quotation marks, and attributes the pair to "Stephen Jay Gould, 1980, The Panda's Thumb."  The attribution is impossible, because the first sentence was published in Discover in May 1981 and The Panda's Thumb appeared in 1980.  The article uses it for a sentence attributed to Gould that appears in no Gould text it has been able to open. ↩ ↩

Snelling, Andrew A.  “Where are all the human fossils?”  Creation 14(1):28–33.  Dec 1991.  “Wise, K.P., 'The Flood and the fossil record', an informal talk given at the Institute for Creation Research, San Diego (USA) on August 17, 1988.”  A magazine article by Andrew Snelling, reprinted online by Answers in Genesis, on why no human fossils have been found in the Flood sediments.  Its figures for the fossil record (95% marine, 0.25% vertebrates, and 1% of those more than a single bone) cite nothing but an informal talk, "The Flood and the fossil record," given by K.P. Wise at the Institute for Creation Research in San Diego on 17 August 1988, in its footnotes 6 and 8.  There is no transcript, no data and no stated method behind that talk, and no peer-reviewed source for the chain exists anywhere.  The article uses it for a statistic repeated for thirty-five years that has no source. ↩ ↩

Berean Archive.  “Collector's curves for fourteen clades.”  Paleobiology Database.  n.d.  Counts of genus names by naming year for fourteen clades, with accepted extinct genera only and an analysis cutoff of 2023, because name entry lags publication.  Ichnotaxa (footprint names) and form taxa (eggshell and isolated-tooth names) are left out, along with junior synonyms and nomina dubia, because they otherwise inflate the dinosaur genus count by about half: Dinosauria returns 3,213 genus records unfiltered and 1,592 filtered.  Taxa the database does not flag extinct are dropped as well, because a group with living members otherwise carries genera named by Linnaeus from living specimens, which back-loads the eighteenth century with names that have nothing to do with fossil discovery.  The article uses the curves to show that dinosaur discovery is still accelerating while trilobite discovery has nearly stopped, and as a ranking of how mature each search is, not as a measure of what is left. ↩ ↩ ↩

1. 13a.

   Restricted by continent, with birds, form taxa and ichnotaxa excluded.  Database counts of accepted, extinct, regular-taxon non-avian dinosaur genera by naming year: 390 by 1990, 564 by 2000, 727 by 2006, 896 by 2010, 1,108 by 2015, 1,331 by 2020, 1,436 by 2023, and 1,534 in total.  New genera per year 1994–2025: 17, 12, 19, 12, 26, 29, 22, 28, 13, 28, 31, 36, 27, 39, 23, 46, 61, 49, 33, 46, 47, 37, 36, 46, 39, 55, 47, 43, 34, 28, 44, 36.  Wang & Dodson's 527 and the database's 727 are not the same count and must not be reported as one series, because Wang and Dodson worked from a stricter hand-vetted list and rejected genera the database marks accepted. ↩ ↩

2. 13b.

   Genus-rank names with no parseable naming year: Foraminifera 1,358 against 503 with a year; Brachiopoda 1,600 against 3,001; Echinodermata 586; Ammonoidea 503; Gastropoda 466; Bivalvia 427.  Trilobita has 62 and Dinosauria has 1.  Unattributed names are, if anything, likelier to be older names entered from compendia, so dropping them makes a curve look more recent-weighted than it is, which pushes against the deceleration conclusion rather than toward it. ↩

3. 13c.

   Run at the species, genus and family ranks, with extinct taxa only and the identical query, filters, cutoff and statistics as the main run.  Naming year is the year the name at that rank was published; for a family that is normally the year of its type genus or later, not the year of its earliest-described member.  Accepted extinct family names, and the number of them published in the last 25 complete years: Insecta 367 of 981; Reptilia 120 of 551; Mammalia 71 of 481; Echinodermata 70 of 537; Actinopterygii 58 of 225; Gastropoda 55 of 191; non-avian Dinosauria 43 of 149; Bivalvia 35 of 225; Brachiopoda 30 of 451; Ammonoidea 26 of 379; Aves 18 of 92; Amphibia 4 of 61; Foraminifera 3 of 129; Trilobita 3 of 182.  New accepted extinct family names by decade for non-avian dinosaurs, 1970s to 2010s: 13, 17, 22, 21, 16, with 449 new genera in the 2010s against 16 new families.  Trilobita, 1950s to 2010s: 55, 14, 5, 10, 4, 2, 1.  Foraminifera, 1980s to 2010s: 41, 8, 1, 1.  Brachiopoda, 1960s and 2010s: 86 and 7.  The forty-year OLS slope of new family names per year is negative in twelve of the fourteen clades; the two positive rows are Actinopterygii at +0.010 and Insecta at +0.298, against a genus-rank slope of +1.110 for non-avian dinosaurs and +4.747 for insects.  Species per family among accepted extinct taxa: non-avian dinosaurs 1,682 ÷ 149 = 11.3, Insecta 27,847 ÷ 981 = 28.4, Trilobita 8,065 ÷ 182 = 44.3.  The genus column reproduces the main run to within one name in two rows, Trilobita 3,375 here against 3,374 there and Brachiopoda 3,002 against 3,001, which is six days of database drift, not a method difference. ↩ ↩ ↩

4. 13d.

   Dropping the extinct-only filter is the control for the two rows where family rank looks less mature than genus rank.  The filter removes any taxon the database does not flag extinct, and a family counts as extant if a single member of it survives, so at family rank it cuts a living-dominated clade to its wholly-extinct residue: Gastropoda 671 accepted family names to 191, Bivalvia 405 to 225, Aves 284 to 92, Mammalia 714 to 481, Insecta 1,953 to 981.  Unfiltered, the share named in the last 25 years, family against genus, is Gastropoda 12.1% against 11.6%, and Bivalvia 9.4% against 12.0%.  So the bivalve exception is entirely a filter artifact and the gastropod one nearly so. ↩

Berean Archive.  “Range gaps in the Paleobiology Database.”  Paleobiology Database.  n.d.  A computed search, by phylum, for genera that disappear from the catalogued record for a stretch and then come back.  Each occurrence is assigned to an ICS stage, and only occurrences whose whole age bracket falls inside a single stage are used, so a loosely dated collection can neither create nor fill a gap.  Of 65,386 distinct genera, 16,804 have occurrences in three or more stages; 1,239,246 occurrences are pinned to one stage and identified to genus, which is 61.9% of the database.  The article uses it because a genus that vanishes and returns shows a demonstrated sampling gap: the lineage was provably there on both sides, so no model or estimate is needed. ↩ ↩ ↩ ↩ ↩

1. 14a.

   By phylum, genera with occurrences in three or more stages and the share with at least one internal gap: Hemichordata 60 genera, 38.3%; Brachiopoda 1,536, 53.6%; Radiozoa 380, 59.7%; Chordata 3,011, 72.2%; Foraminifera 731, 74.8%; Cnidaria 837, 77.7%; Arthropoda 2,106, 78.8%; Echinodermata 543, 78.8%; Mollusca 4,807, 80.1%; Porifera 360, 81.7%; Bryozoa 450, 86.7%; Pteridophyta 179, 93.3%; Angiospermae 385, 93.8%.  Gap length in stages: 43.3% span a single stage, 19.6% two, 11.8% three, and 6.1% (1,790 of 29,587) span ten or more.  An earlier version of the article printed 4.3% for the ten-or-more share; that could not be reproduced on re-derivation and appears to have been the five-stage row, which is 4.3% on this run. ↩

Berean Archive.  “Sister-taxon stratigraphic gaps.”  Paleobiology Database.  n.d.  A computed measure of how long a taxon's first appearance trails its nearest sister, across the animal kingdom (Metazoa) at every rank from species to phylum.  Every accepted taxon inside Metazoa, leaving out form taxa and ichnotaxa, is grouped by its immediate parent; for each parent with two or more dated children, the gap runs from the oldest child's first appearance to the second oldest's, with every age bracket taken in the direction least favourable to a gap, and the result floored at zero.  "Sister" here means "shares an immediate parent in the database," which is a classification relationship, not a phylogenetic one.  The article uses it because nobody in this dispute appears to have measured the stratigraphic gap at each rank. ↩ ↩ ↩

Berean Archive.  “Sampling and geographic bias in the Paleobiology Database.”  Paleobiology Database.  n.d.  A count of all 277,490 collections by country and by phylum, run to measure how lopsided the sampling is.  The United States holds 62,063 (22.37%), China 15,314 (5.52%), Canada 14,895 (5.37%), an ocean-basin code 12,354 (4.45%), Russia 11,481 (4.14%), the United Kingdom 9,158 (3.30%), Australia 8,723 (3.14%), France 8,489 (3.06%), Spain 6,881 (2.48%) and Germany 6,114 (2.20%).  The top three hold 33.26%, the top five 41.85%, the top ten 56.91% and the top twenty 72.96%, and the Gini coefficient across 255 country codes is 0.861.  By phylum, five phyla hold 68.3% of all occurrences: Mollusca 502,370 (25.11%), Chordata 313,680 (15.68%), Foraminifera 196,749 (9.83%), Brachiopoda 187,542 (9.37%) and Arthropoda 167,040 (8.35%).  Only 47.8% of occurrences are resolved to species and 41.6% to genus, so a claim about ordering at genus rank rests on 89.4% of the database.  The article uses it because every number that depends on the database has to be discounted for this bias. ↩ ↩ ↩

1. 16a.

   Over the 99 ICS stages with at least ten stage-pinned collections, Pearson r(collections, genera) = 0.645; Spearman ρ(collections, genera) = 0.757; Pearson r on log10 of both = 0.851; and Spearman ρ(occurrences, genera) = 0.888.  Sampling density spans four orders of magnitude, from 21 occurrences per million years in the earliest Cambrian to 305,013 per million years in the Late Pleistocene. ↩ ↩ ↩

Berean Archive.  “Fossil coverage of living families.”  Paleobiology Database and GBIF backbone taxonomy.  n.d.  A computed count of how many living families have any fossil record.  The denominator is the living families in the GBIF backbone taxonomy, not the database's own list of living families, because a living family enters a paleontological database largely because somebody wrote about its fossils; asking what share of the database's own extant families have occurrences is circular and biased high.  Families are matched by name, so a name mismatch counts as no fossil record and the resulting figures are lower bounds.  The article uses it for the completeness statistic most often quoted in the debate and almost never computed. ↩ ↩

Hunt, Gene.  “The relative importance of directional change, random walks, and stasis in the evolution of fossil lineages.”  Proceedings of the National Academy of Sciences 104(47):18404–18408.  Edited by Michal Kucera.  2007.  “An important advantage of this model-based approach is that no evolutionary mode is granted privileged null status; all models are compared on an equal footing based on their empirical support.”  Table 1:  Table 1 supplies the counts that the article tabulates.  A statistical survey of more than 250 evolving traits in fossil lineages, each fitted to three models: directional change, a random walk and stasis.  Directional change was the best-supported model in only 5% of them, and the rest split nearly equally between random walks and stasis.  The article uses it as the best single study of how fossil lineages change. ↩

1. 18a.

   Pages 18406–18407:  Hunt warns against stretching the word stasis to cover every pattern that is not strongly directional.  “At times, the term stasis has been applied more widely to describe any evolutionary pattern that is not strongly directional.  Such an overly broad definition is at odds with most recent accounts of stasis (9, 23) and obscures real differences between the fluctuating but stationary trajectories of stasis and the unbounded meandering of random walks.” ↩

2. 18b.

   Page 18406:  Hunt says that even the 5% figure is probably too high.  “because researchers have preferentially studied traits and lineages with prior evidence for gradual change (9, 16), the 5% incidence of directionality is likely to be an overestimate.” ↩

3. 18c.

   Page 18405:  Duration: G = 2.43, df = 4, P = 0.66.  Resolution: G = 4.77, df = 4, P = 0.31.  The resolution bins were cut at 0.08 and 0.34 million years.  Trait type: stasis wins in about 60% of shape traits against about 37% of size traits, G = 11.59, df = 2, P = 0.003.  Hunt draws the inference himself: a population-structure explanation of stasis should affect all traits equally, so a trait-type difference points at something intrinsic to the trait. ↩ ↩ ↩ ↩

4. 18d.

   “On average, the best-supported model accounts for 77% of the available Akaike weight … in only 20% of sequences can both suboptimal models be confidently dismissed.” ↩

5. 18e.

   Table 2:  Table 2 gives, for each group, the sequences, lineages and studies, followed by the median number of sampling horizons: planktonic foraminifera 23 (8, 6), median 48; radiolaria 9 (3, 1), median 9; conodonts 9 (1, 1), median 6; benthic foraminifera 37 (5, 4), median 19; ostracodes 60 (12, 3), median 7; molluscs 70 (10, 6), median 10; mammals 40 (11, 7), median 7; teleost fish 2 (2, 2), median 17.5; trilobites 1 (1, 1), median 22.  So 138 of 251 sequences are microfossils and 41 are planktic.  Table 1:  Table 1 gives the model counts, in the order planktic microfossil, benthic microfossil, macrofossil: directional 13 (5, 3, 5), random walk 123 (24, 57, 42) and stasis 115 (12, 37, 66).  Page 18405:  Hunt's own statement of the pattern is:  “Directional sequences are more common in planktonic microfossils than in either benthic microfossils (12% vs. 3%; G = 3.92, df = 1, P = 0.048) or benthic macrofossils (12% vs. 2%; G = 4.30, df = 1, P = 0.038).” ↩ ↩ ↩

6. 18f.

   Page 18406:  The sequence is shell shape in Contusotruncana from DSDP Site 384, from Kucera & Malmgren 1998.  “yet the parameter estimates of this model imply only a trivial excess of increases on generational time scales:  On average, this trait should increase in only 50.1% of years.”  Hunt adds that the net rate.  “is substantially slower than the expectation from mutation–drift balance.”  Michal Kucera, who edited the paper, is also the first author of the 1998 study that Hunt uses as his directional exemplar. ↩

7. 18g.

   Page 18406:  Hunt says that even the most promising examples do not support sustained directional change, although single interval punctuations are sometimes implied.  “I have used these methods to analyze many fossil sequences, but even the most promising examples (e.g., refs. 21 and 22) do not support models of sustained directional change, although single interval punctuations are sometimes implied (unpublished data).”  Reference 21 is Malmgren & Kennett 1981, Paleobiology 7:230–240, on the Globorotalia conoidea–inflata lineage, the flagship planktic foraminiferal gradualism case.  Reference 22 is Kellogg 1975, Paleobiology 1:359–370, on Pseudocubus vema, the flagship radiolarian case.  Both are planktic microfossil lineages. ↩

Hunt, Gene et al.  “Simple versus complex models of trait evolution and stasis as a response to environmental change.”  Proceedings of the National Academy of Sciences 112(16):4885–4890.  2015.  A database of 709 sequences from nearly 200 fossil lineages, built by merging Hunt 2007 with Hopkins & Lidgard 2012 and adding new studies, with sequences running from 7 to 114 samples (median 14), each fitted with nine models instead of the three used in 2007.  The sequences are 249 planktic microfossils (median 22 samples), 257 invertebrate macrofossils (median 14), 112 vertebrates (median 8) and 91 benthic microfossils (median 14); microfossils are 340 of the 709, or 48%, down from 55% in 2007 but still nearly half, so the canonical modern test of stasis against gradualism is run about half on plankton and shelled protists, on both databases.  The authors find stasis common and directional change in about 13% of sequences, and the article uses the paper because its authors reject the 2007 finding that plankton change directionally more often than other fossils. ↩

1. 19a.

   Pages 4886–4887:  “Previous indications that directional trends are more likely to be observed in planktonic microfossils than in other fossil groups (14, 29, 30) are not supported by logistic regression (SI Appendix, Table S3).”  Reference 14 in that sentence is Hunt 2007, the paper whose plankton result the sentence rejects.  References 29 and 30 are Jackson & Cheetham 1999 and McKinney 1985, the same two sources that Hunt 2007 cited for the "geological and biological grounds" for expecting plankton to change gradually. ↩ ↩

2. 19b.

   Page 4887:  “Similarly, the raw proportion of sequences best fit by directional models is not very different among planktonic microfossils (15.3%), benthic microfossils (19.8%), and vertebrates (14.2%).  This proportion is notably lower among macroinvertebrates (7.8%), but this figure may be depressed by two studies (31, 32) that included numerous traits in which stasis dominated.”  Two cautions travel with these figures.  They do not measure quite the same quantity as Hunt 2007's: the 2015 counts, including mode-shift models with a directional component, use full maximum likelihood rather than REML (restricted maximum likelihood).  Both choices raise the directional count, which makes the reversal more striking, since the more sensitive method still finds no plankton advantage.  Refitting the deposited sequences gives 14.9%, 17.6%, 14.3% and 8.9% for these four groups.  Page 4886:  “Both kinds of stasis are more prevalent in marine compared with terrestrial or lacustrine environments, and in macroinvertebrates and vertebrates compared with microfossils (SI Appendix, Tables S1 and S2).”  One part of the 2007 contrast does persist, and it is printed rather than hidden.  What collapsed is not "macrofossils show more stasis"; it is the claim that the difference is made up by directional evolution in plankton. ↩ ↩

3. 19c.

   Page 4886:  “Strict stasis is less probable with increasing temporal duration of sequences, and the broad version of stasis is also less likely to be best supported as the number of samples increases.”  “Complex models are substantially more likely to be favored with increasing numbers of samples, but not with increasing temporal duration (Table 2)”.  “For example, 94% of sequences (58 of 62) with >40 samples are best supported by a model that implies complex evolutionary dynamics.”  The complex models are punctuation and the four mode shifts, which take 175 of the 709 sequences: 25% overall, but 45% of the 388 sequences long enough to be fitted with one at all.  Directional models take 13% of the 709.  Page 4887:  The authors' own reading is:  “Instead, the ability of complex models to win is limited only by the information content of a sequence, and therefore its number of samples.”  The sentence cuts both ways.  Every simple model, stasis included, is partly an artefact of thin data.  The messiness that better data reveals is, for this article, discontinuous rather than gradual. ↩ ↩ ↩ ↩

4. 19d.

   Page 4886:  “In addition, age model error that is caused by undetected variation in sedimentation rate has a tendency to spuriously favor complex evolutionary models, but the effect is not large enough to account for these patterns (SI Appendix, SI Text).”  Undetected changes in how fast sediment piled up can make a sequence look punctuated, because they stretch and compress the time axis.  The authors modelled that effect and found it too small to explain the pattern.  It is the most important methodological check in this literature for this article's purposes, and it was run by authors with no sympathy for its conclusions. ↩ ↩

Hunt, Gene et al.  “Data from: Simple versus complex models of trait evolution and stasis as a response to environmental change.”  Dryad Digital Repository.  2015.  The authors' deposit of 771 ancestor-descendant trait series from the fossil record, with a metadata table carrying lineage, taxon, trait type, environment, source study and sample count, and a list of source publications; the article analysed the 709 series with seven or more samples.  The authors printed four percentages and a set of regressions but never published the table behind them, so the article refitted all 709 sequences from this deposit and printed that table. ↩ ↩

1. 20a.

   Validation against Hunt et al.'s Table 1, published count then refit count: strict stasis 124 / 125, stasis 147 / 125, random walk 201 / 184, directional 62 / 60, punctuation 67 / 111, stasis-to-random-walk 19 / 22, stasis-to-directional 16 / 16, random-walk-to-stasis 59 / 50, directional-to-stasis 14 / 16.  Directional models combined, 92 / 92.  Stasis of either kind, 271 / 250.  Complex models, 175 / 215.  The punctuation model is the one real discrepancy.  It gains 44 sequences, taken almost entirely from broad stasis and the random walk, and it is the reason the complex share reads 30.3% here against 24.7% in the paper.  The likely cause is in the package's own changelog, which records an "improved default setting for all optimization functions that improves convergence behavior" in version 0.6.1, along with standardised bounds handling in L-BFGS-B.  Punctuation has the most parameters to place and the most breakpoints to scan, so a poorly converging optimiser loses with it most, and better convergence should raise its win rate.  That is a hypothesis about a version difference, not a finding.  The direction matters for honesty: the drift runs against this article, since it moves sequences out of stasis. ↩ ↩

2. 20b.

   Stasis wins is the share of a clade's sequences whose best-supported model is strict stasis or broad stasis.  Directional wins is the share taking the pure directional model or either mode-shift model with a directional phase.  Clade labels are the deposit's own taxon field, with its two spellings of planktic foraminifera merged and its six nannolith rows folded into coccolithophores.  Six further clades are in the database and are too thin to print: birds 8 sequences, echinoids 8, fish 7, brachiopods 6, gastropods 2 and hemichordates 1.  They are in the released table.  The strict-stasis share sharpens the same contrast: bryozoans 50.0% and trilobites 63.6% against planktic foraminifera 4.2% and diatoms 0.0%. ↩

3. 20c.

   The full model breakdown of the top row, which is the 62 sequences with more than 40 samples, is punctuation 32.3%, random-walk-to-stasis 35.5%, stasis-to-random-walk 11.3%, directional-to-stasis 9.7%, stasis-to-directional 6.5%, pure directional 3.2%, unbiased random walk 1.6%, stasis 0.0% and strict stasis 0.0%.  The structural caveat is real: the four mode-shift models and the punctuation model require at least 14 samples, so the 321 sequences in the thinnest bin were never offered a complex model, and their 45.5% stasis is a win over three rivals rather than eight.  The comparison this does not affect is between the 21–40 bin and the bin above it: both were offered all nine models, and stasis falls from 46.5% to 8.7% between them. ↩ ↩ ↩ ↩

4. 20d.

   Logistic regression of stasis (strict or broad) on group alone finds every group significantly more static than planktic microfossils: benthic microfossils P = 0.013, vertebrates P = 2.3 × 10⁻⁶, macroinvertebrates P = 2.7 × 10⁻¹⁵, likelihood-ratio P = 9.4 × 10⁻¹⁶.  Adding log sample count changes the picture: benthic microfossils P = 0.54 and vertebrates P = 0.36 both fall away, macroinvertebrates survive at P = 6.2 × 10⁻⁷, and the sample-count term is the strongest in the model, at −1.50, P = 6.4 × 10⁻¹³.  Regressing a directional outcome on group alone gives a likelihood-ratio P of 0.087, which is this article's own independent confirmation of a statement Hunt et al. supported with a supplementary table.  Hunt et al.'s Table 2 is reproduced closely with their own specification, stepwise AIC over sample count, duration, environment and taxon category.  Published, then refit: strict stasis, dur(−), env, taxon, against ns(−), dur(−), env, taxon; strict-plus-broad stasis, ns(−), dur(−), env, taxon, against the same four; complex, ns(+), env, taxon, against the same three.  The directional row is the one needing care.  Hunt et al. report dur(+), taxon.  The pure directional model refit here retains duration alone, positive, P = 0.0016, which is their result.  Fitting all directional-containing models keeps sample count as well, positive, P = 0.0026.  So "directionality tracks duration, not sampling" is right for sustained trends and incomplete for the 13% figure, which counts partial ones. ↩ ↩

5. 20e.

   Hunt et al.'s references 31 and 32 are Kelley 1983 on eight Chesapeake Group molluscs and Cheetham, Sanner and Jackson 2007 on Metrarabdotos bryozoans.  They contribute 44 and 72 sequences, which is 116 of the 257 macroinvertebrate sequences and 116 of all 709.  Macroinvertebrates, all 257: strict stasis 28.4%, stasis 52.1%, directional 8.9%.  The 141 that remain when those two studies are removed: strict stasis 17.0%, stasis 30.5%, directional 14.2%.  The Cheetham bryozoans on their own, 72 sequences: strict stasis 54.2%, stasis 83.3%, directional 0.0%.  In the 2007 three-model currency, the same removal takes macroinvertebrate stasis from 64.6% to 44.0%.  This cuts two ways.  It removes the last apparent group difference in directionality, which supports Hunt et al.  It also means the most static block in the database is one bryozoan monograph, so no group-level statement about macroinvertebrates in this literature is independent of it.  The database holds 116 distinct source studies for 709 sequences, and the largest single study supplies 10.2% of them, so treating sequences as independent observations overstates the sample throughout, for the authors and for this article alike. ↩ ↩

6. 20f.

   Three models only are fitted here: directional, unbiased random walk and stasis, by restricted maximum likelihood on ancestor-to-descendant differences.  That is the method of Hunt 2007, available in paleoTS as the "AD" option.  All 709 sequences under the 2007 method give directional 39 sequences (5.5%), stasis 352 (49.6%) and random walk 318 (44.9%), against Hunt 2007's 5%, 46% and 49% on 251 sequences.  By group, the directional share is planktic microfossils 6.4%, benthic microfossils 3.3%, macroinvertebrates 4.7% and vertebrates 7.1%.  The 2007 contrast that was withdrawn had been 12.2% planktic against 3.1% benthic and 4.4% macro.  The plankton advantage more than halves with the method held fixed and the database tripled, and vertebrates end up on top.  The trait-type result of 2007 survives on the larger database but weakens sharply: stasis wins 41.5% of shape traits against 33.4% of size traits, χ² = 4.02, P = 0.045, where Hunt 2007 had about 60% against about 37% at P = 0.003. ↩ ↩

Renaudie, Johan, and David B. Lazarus.  “Cenozoic pelagic accumulation rates and biased sampling of the deep-sea record.”  Biogeosciences 22:1929–1946.  2025.  Page 1938:  “Age model hiatuses (> 0.5 Myr) in NSB deep-sea sediment sections are relatively rare (median of 0, mean of 0.93 hiatuses per section; Fig. 3).  Most are of fairly short duration (median 2.98 Myr), though there is an extended tail of rare, longer-duration hiatuses, thus a mean duration of 6.47 Myr (Fig. 3).  Overall, the section age models are on average > 85 % complete (fraction of time represented by sedimentation) (Fig. 3).”  Both authors built the NSB deep-sea microfossil database, so this is that community measuring its own archive.  The article uses it for the hiatus figures of the deep-sea record.  Early drilling counts had much higher hiatus frequencies, but the authors supersede them and give the reasons: early holes were sited at poorly known locations, later ones were placed using high-resolution seismics, and the early counts treated a missing biostratigraphic zone as missing time, which over-counts, because zones go unrecognised for biogeographic, preservational and taxonomic reasons that have nothing to do with a gap.  The article does not use those older figures.  Hiatus frequency, the share of sections spanning an interval that fail to record it, runs about 22–24% from the Eocene to the Middle Miocene and rises to about 30% in the early Paleocene. ↩ ↩

1. 21a.

   Page 1938:  “Crucially, there is no trend in by-section SAR over geologic time when compiled for the entire dataset (Fig. 5):  the slope of a linear regression is insignificantly different from zero, and the R 2 is also essentially zero (1.277 × 10−5).  Variations in rate within sections thus are modest and essentially random vs. time.”  SAR is sediment accumulation rate.  This bears directly on Part 1 of the article: if accumulation rate wandered within a core, every apparent change of evolutionary rate in a deep-sea sequence would be suspect, because a compressed interval looks like a fast change. ↩ ↩

2. 21b.

   Abstract:  “Here we show that the Cenozoic record of sedimentation recovered by deep-sea drilling has a strong, systematic bias towards lower rates of sedimentation with increasing age.  When this bias is removed, accumulation rates are shown to actually decline by ca. 2 times over the Cenozoic.”  Page 1938:  Measured rates are:  “SARs for individual age model segments in the NSB library vary dramatically (< 0.5 to > 30 cm kyr−1 range; interquartile ratio of 4.43; Fig. 5).”  The binned means fall from about 3.9 cm/kyr in the youngest bin to under 0.7 in the oldest, the six-fold apparent decline that is the artefact.  Page 1941:  Their own generalisation:  “Although we have not explored the implications of bias beyond that of drilling pelagic sediments, it is to be expected that a similar effect exists in other types of drilled geologic sections and even to some degree ordinary land sections.”  This is a "the record gets worse with age for reasons that have nothing to do with what lived" result, published in 2025 by mainstream workers about the cleanest archive there is. ↩ ↩

Kucera, Michal, and Björn A. Malmgren.  “Differences between evolution of mean form and evolution of new morphotypes: an example from Late Cretaceous planktonic foraminifera.”  Paleobiology 24(1):49–63.  1998.  “Therefore, despite the gradual change in 'mean form,' the morphological evolution in the Contusotruncana lineage differs from the classical image of phyletic gradualism.”  The authors give this as the reason for the gradual increase in mean shell conicity:  “arose from the combination of a rapid development of highly conical shells after 68.5 Ma and a long-term trend of progressive disappearance of the ancestral morphotype.”  The authors separated traits controlled by geographical variation from those that changed at the same time in geographically distant populations, and listed shell size with the first group:  “controlled by geographical variation (size, spiral roundness of the test, percentage of kummerform specimens)”.  A measurement study of the Contusotruncana lineage of planktonic foraminifera, floating single-celled sea animals with shells, from the Late Cretaceous.  The authors found that the average shell form changed gradually, while highly conical shells appeared quickly and then slowly replaced the ancestral form.  The article uses it because these are the measurements behind Hunt 2007's Fig. 1A, the strongest directional sequence in Hunt's database, and because the smooth average hides a change in the mix of forms. ↩ ↩

Erwin, Douglas H., and Robert L. Anstey.  “Speciation in the fossil record.”  New Approaches to Speciation in the Fossil Record, 11–38.  Columbia University Press, New York.  1995.  The opening chapter of a book of papers on speciation in the fossil record.  The article uses it because its authors tallied 58 published lineage studies from the previous 22 years and concluded that no single pattern appears dominant. ↩

1. 23a.

   Page 38:  “The results so far ostensibly support a pluralism of tempos and modes in the historical record of speciation.  No single pattern appears dominant, and their variety suggests that new research agendas are needed to investigate the conditions under which any one of them might be predicted.”  Page 23:  The book's closing summary says the same, that speciation.  “is sometimes gradual and sometimes punctuated, and … no one mode characterizes this very complicated process.” ↩

2. 23b.

   Page 42:  “An interesting departure in table 1.1 from conventional expectations is that fully one-fourth of the studies we assessed describe single lineages that sequentially or simultaneously display both gradualism and stasis — more than half of which represent metazoans.” ↩

3. 23c.

   Page 44:  “More than one-fourth of the studies in table 1.1 are based on protozoans, which may not form species comparable to those in outbreeding metazoans.”  Page 39:  “There is simply no such thing as complete stasis in features, such as shapes, measured on a continuous scale.” ↩

4. 23d.

   Page 43:  “all of the taxa displaying gradualism in the particular characters studied by the researchers in this table must simultaneously display stasis in others, namely those characters that define and delineate their lineage.”  Their example is the peppered moth, which.  “displays gradualism only in its protective coloration, and ostensibly stasis in all other features.” ↩

5. 23e.

   Pages 42, 44:  On Ozawa's Lepidolina multiseptata, the lineage.  “evolved gradually in four characters, but not in five others; therefore they might also be moved to the center column.”  They add that the trend.  “might likewise reflect an extremely sluggish decrease in outbreeding during the Permian.” ↩

Prothero, Donald R.  “Does climatic change drive mammalian evolution?”  GSA Today 9(9):1–7.  Geological Society of America.  Sep 1999.  “Of 70 species known from the earliest Orellan, 62 persisted unchanged into the late Orellan.”  The Orellan is a span of the early Oligocene.  On the exceptions, a few archaic Chadronian groups (brontotheres, oromerycid artiodactyls, cylindrodont rodents) disappeared, and the oreodont Miniochoerus dwarfed slightly.  “but most mammalian lineages showed no changes worth documenting.  Compared to the 177 species now documented for this interval, this is a remarkably mild response.”  A short article in GSA Today, the monthly news magazine of the Geological Society of America.  Donald Prothero, a geologist at Occidental College, compared fossil mammals before and during four of the largest climate changes of the Cenozoic, the geological era since the dinosaurs died out, and found little response in most species.  The article uses it as its most useful single-fauna case, since its author is a critic of creationism working on an unusually complete fossil record. ↩

1. 24a.

   “This lack of change cannot be dismissed as an artifact of sampling or preservation, because the White River Group in eastern Wyoming is densely fossiliferous through all of the relevant interval.” ↩

2. 24b.

   “the striking feature of each of these abrupt climatic changes is the lack of response of land mammals, even though in each example, it is clear that land plants and other elements of the terrestrial biota are responding.”  The four events tested are the late-middle Eocene cooling (~37 Ma), the Eocene–Oligocene transition (~33 Ma), the late Miocene grassland expansion (~7 Ma), and the Plio-Pleistocene glacial cycles.  The underlying monograph is Prothero & Heaton, "Faunal stability during the early Oligocene climatic crash," Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996): 257–283. ↩

Prothero, Donald R.  “Darwin's Legacy.”  eSkeptic.  Feb 15, 2012.  The figure he recalls is.  “over 160 well-dated, well-sampled lineages”.  “No matter how many presentations I give where I show these data, no one (including myself) has a good explanation yet for such widespread stasis despite the obvious selective pressures of changing climate.”  A column in eSkeptic in which Donald Prothero, recalling his own 1982 Columbia doctoral dissertation, gives a figure.  Prothero is a militant anti-creationist, and he spent the weekend before writing the column visiting a creation museum to mock it, which is exactly what makes him a useful witness.  The article uses the column for his admission that stasis, a species staying unchanged for long periods, is still unexplained under changing climate, a point the article says cannot be dismissed as a creationist gloss. ↩ ↩

Uyeda, Josef C. et al.  “The million-year wait for macroevolutionary bursts.”  Proceedings of the National Academy of Sciences 108(38):15908–15913.  2011.  “we find a complete absence of a time-span effect up to ∼1 Myr, but this pattern then yields to a pattern of increasing divergence on timescales above 1–10 Myr.”  The segmented regression breakpoint is at 10^4.82 years, about 66,000.  The best-fitting model was multiple-burst, with an expected waiting time to displacement of the adaptive optimum above 10 Myr and an ancestor-to-descendant size ratio per burst of about 1.28.  A research paper that pools more than 8,000 body-size measurements of mammals, lizards and snakes, and birds, spanning 0.2 years to 357 million years and drawn from field studies, historical records, fossil series, and family trees of living species, then fits mathematical models with random chance built in.  The article uses it to show where stasis (long spells with little change) ends: change does not accumulate up to about a million years, and divergence grows after that. ↩ ↩

1. 26a.

   On the data-source objection, the paper says:  “A segmented regression with a breakpoint at 66,000 y has a lower Akaike's information criterion than a model in which independent regressions are fitted to each of the three major subsets.”  On climate, the paper says:  “substantial changes in global climate over timescales of ≪1 Myr appear to result in bounded divergence characteristic of microevolutionary timescales, rather than dramatically hastening phenotypic evolution.” ↩ ↩

2. 26b.

   “We conclude that some unspecified process of bounded evolution is responsible for the band we observe in our data (Fig. 1).”  They list the models ruled out by Estes and Arnold and add:  “All of these models fail either because they require unrealistic values for microevolutionary processes or because bounded evolution is difficult to achieve under any set of parameter values.” ↩ ↩

Gould, Stephen Jay, and Niles Eldredge.  “Punctuated equilibrium comes of age.”  Nature 366:223–227.  1993.  “its most important implications remain the recognition of stasis as a meaningful and predominant pattern within the history of species, and in the recasting of macroevolution as the differential success of certain species (and their descendants) within clades.”  A review by the two authors who proposed punctuated equilibrium, written 21 years after the idea was first put forward.  They argue that its most lasting contribution is the recognition that stasis, a species persisting without change, is a meaningful and predominant pattern in the history of species.  The article uses it because the authors concede that gradual change is well documented and that their own causal claim has not been validated. ↩

1. 27a.

   “Like all major theories in the sciences of natural history, including natural selection itself, punctuated equilibrium is a claim about relative frequency, not exclusivity.  Phyletic gradualism has been well documented, again across all taxa from microfossils to mammals.”  The authors' own 1993 summary, two sentences later, says:  “The issue remains unsettled.” ↩

2. 27b.

   Page 227:  “But continuing unhappiness, justified this time, focuses upon claims that speciation causes significant morphological change, for no validation of such a position has emerged (while the frequency and efficacy of our original supporting notion, Mayr's 'genetic revolution' in peripheral isolates, has been questioned).” ↩ ↩

3. 27c.

   Page 223:  On speciation, the same paper says:  “for small populations speciating away from a central mass in tens or hundreds of thousands of years, will translate in almost every geological circumstance as a punctuation on a bedding plane, not gradual change up a hill of sediment.” ↩

4. 27d.

   Adopting Futuyma's account and quoting him (Douglas J. Futuyma, American Naturalist 130 [1987]: 465–473):  “In the absence of reproductive isolation, differentiation is broken down by recombination. …  Although speciation does not accelerate evolution within populations, it provides morphological changes with enough permanence to be registered in the fossil record.”  Uyeda et al. 2011 call this.  “a particularly elegant explanation”.  and note that their million-year figure is consistent with it. ↩

Sheldon, Peter R.  “Parallel gradualistic evolution of Ordovician trilobites.”  Nature 330:561–563.  1987.  Hunt's wording is.  “celebrated examples … of gradualism”.  A paper on eight trilobite lineages, an extinct group of sea arthropods, from central Wales, tracked over about three million years, all showing gradual change in rib number in parallel.  Sheldon later argued the counter-intuitive position (Nature 345 [1990]: 772) that gradualism should prevail in stable environments and punctuation in unstable ones, a view that Gould and Eldredge (1993) cite respectfully.  The article uses it because Hunt (2007) counts it among the celebrated examples of gradualism that he could not include, since the underlying measurements were never published in usable form. ↩

Malmgren, B. A. et al.  “Evidence for punctuated gradualism in the Late Neogene Globorotalia tumida lineage of planktonic foraminifera.”  Paleobiology 9(4):377–389.  1983.  “Thus it would appear that the G. tumida lineage, while remaining in relative stasis over a considerable part of its total duration underwent periodic, relatively rapid, morphologic change that did not lead to lineage branching.”  They state what disqualified it as punctuated equilibrium: the transition.  “was not rapid enough (<1% of the descendant species' duration according to definition)”.  and no branching occurred.  Both disqualifications were overturned twenty-six years later by Hull and Norris (2009).  A study of Globorotalia tumida, a lineage of planktonic foraminifera (tiny shelled marine animals) studied in deep-sea sediment cores, which coined the term punctuated gradualism for it.  The authors describe a minor fluctuation with no net change from 10.4 to 5.6 Myr (million years ago), then a transformation lasting about 0.6 Myr, then no major modification for 5 Myr, with sampling every 5–15 kyr (thousands of years) across the transition.  The article uses it because it coined the term and is the case that later work took apart. ↩

Bell, Michael A.  Gasterosteus doryssus sequence from the Miocene Truckee Formation, Nevada.  n.d.  Hunt's Table 2 records the two teleost sequences with a median of 17.5 sampling horizons.  The re-analysis with the same model-fitting machinery is Hunt, Bell & Travis, Evolution 62 (2008): 700–710.  A fossil sequence of the stickleback Gasterosteus doryssus from the Miocene Truckee Formation of Nevada, a varved diatomite, which is a rock laid down in annual layers.  It shows directional change in pelvic and dorsal-spine structures over roughly 20,000 years, at annual resolution.  The article uses it as the highest-resolution fossil vertebrate sequence known, and does not omit it. ↩

Hull, Pincelli M., and Richard D. Norris.  “Evidence for abrupt speciation in a classic case of gradual evolution.”  Proceedings of the National Academy of Sciences 106(50):21224–21229.  Dec 15, 2009.  On the size confound, the paper says:  “In Malmgren et al.'s study (33), a 3-fold increase in mean size accompanies the morphological transition in the G. tumida–G. plesiotumida lineage and is strongly correlated with mean morphology along eigenaxis 2 (EA 2, Pearson's r = 0.89 and P < 0.001).”  Hull and Norris removed the confound by sampling a narrow size fraction, 250–500 µm instead of about 150–500, after which the correlation between size and their own shape axis fell to r = 0.23.  A research paper that re-measures a classic lineage of planktonic foraminifera, tiny shelled marine animals, using 1,140 individuals from 38 depth intervals in a deep-sea core from ODP Site 806B on the Ontong Java Plateau.  The authors measured shape with semilandmark thin-plate-spline morphometrics, a newer method that places points along each outline, instead of the eigenshape method of the 1983 study.  The article uses it because it shows that the classic case of gradual change was not what it was reported to be. ↩ ↩ ↩

1. 31a.

   On the third form, the paper says the flattened, almost exclusively dextrally coiled morphotype.  “becomes abruptly dominant in our record at 6.225 Ma, and persists for 414,000 years until 5.819 Ma.”  It differs from both named species along the axis carrying 53% of morphological variance, and coiling direction marks cryptic species in two living planktic foraminifera, as a 2001 study of them reports.  On the transition, the paper says:  “It is clear that the first G. tumida appeared very rapidly, in less than 45,000 years, rather than the hundreds of thousands of years inferred from previous work.”  The shift exceeds 1.5 standard deviations inside a 44,000-year span, with pre- and post-shift forms co-occurring at 5.819 Ma and population variance elevated across it, the signature of two populations overlapping rather than one transforming. ↩ ↩

2. 31b.

   On the limits of morphometric studies, the paper says:  “Consequently, this and other morphometric studies may lack adequate information to clearly differentiate speciation from within lineage evolution.”  On the older method, it says:  “It is disturbing that eigenshape analysis (used by previous authors) fails to detect differences in compression of individuals in the equatorial Pacific, where both semilandmark TPS analysis and visual inspection confirm the distinct morphology of dextrally coiled individuals.”  A roughly 1.4-fold rise in mean centroid size survives their size control.  Their mixture-model test for two overlapping sinistral populations has.  “very low power (0.01–0.48)”.  and cannot resolve whether ancestor and descendant co-occur.  The authors allow that the sinistral series may be.  “in possible agreement with Malmgren et al.'s interpretation of gradual, within-lineage evolution.”  The paper shows that the classic case was not what it was reported to be.  It does not prove punctuation. ↩ ↩

Bijma, Jelle et al.  “Lunar and semi-lunar reproductive cycles in some spinose planktonic foraminifers.”  Journal of Foraminiferal Research 20(2):117–127.  1990.  Page 117:  “Based on a time series of plankton-net tows, a synodic lunar reproductive cycle is demonstrated in Globigerinoides sacculifer (Brady). The reproductive cycles of Globigerinoides ruber (d'Orbigny) and probably Globigerinella siphonifera (d'Orbigny) are characterized by a semi-lunar periodicity.”  The arithmetic is the article's: a synodic lunar cycle is 29.5 days and a semi-lunar cycle about 14.8, so roughly 12 to 25 generations a year.  The three species are spinose surface dwellers, while the lineages in the article's first part are mostly non-spinose deep-dwelling globorotaliids, whose reproductive periodicity is not established by this work and whose cycle could not be found in any source.  The caveat matters, and the body text depends on it.  If globorotaliids are annual, dividing the generation counts by 12 to 25 still leaves tens of thousands of generations in a single sample.  The conclusion is not sensitive to that assumption, but the exact number is.  “Inter-species differences in lifespan result in a different empty shell output. As a consequence, the sediment does not reflect the situation in the water column.”  One further sentence in the same abstract bears on the article's first part.  A field study that towed plankton nets over time to record the breeding cycles of three spinose planktonic foraminifers, which are single-celled shelled plankton.  Its findings are a synodic lunar cycle in one species and a semi-lunar periodicity in two others, one of them probably.  The article uses it for the number of generations these plankton pass through in a year, which it puts at roughly 12 to 25. ↩ ↩

Norris, Richard D. et al.  “What is gradualism? Cryptic speciation in globorotaliid foraminifera.”  Paleobiology 22(3):386–405.  1996.  “The silhouette of fohsellid shells displays a nearly unbroken anagenetic trend, yet isotopic data show that the fohsellids changed their depth of reproduction during the anagenetic evolution of their skeletons.”  “Although anagenetic trends have been widely cited as evidence for gradual speciation in planktic foraminifera and other microfossil groups, our data suggest that they should not always be considered to record either the tempo or mode of speciation.”  The authors propose a testable mechanism: shell shape is set during the main growth phase near the surface, while reproductive isolation is established at depth at the end of the life cycle, so the two can decouple.  A study that added stable isotopes (chemical ratios in the shell) to the fohsellid foraminifera, a planktonic group whose shell shape changes in an unbroken, smooth series.  The authors found that the animals changed their depth of reproduction beneath that smooth trend.  The article uses it because the fohsellids are the textbook case of unbroken gradual change, and this paper shows that such a trend does not by itself reveal how speciation happened. ↩

Wei, Kuo-Yen, and James P. Kennett.  “Phyletic gradualism and punctuated equilibrium in the late Neogene planktonic foraminiferal clade Globoconella.”  Paleobiology 14(4):345–363.  1988.  “During the Late Miocene (7 Ma to 5 Ma), populations of the ancestral species Globorotalia (Globoconella) conomiozea formed a geographic cline showing continuous morphological variation from the temperate sites (DSDP 284 and 207A) to the warm subtropical sites (DSDP 208 and 588).”  The central temperate populations transformed gradually over about 0.2 Myr, while the peripheral subtropical populations produced G. pliozea.  “very rapidly within an interval of less than 0.01 m.y.”  and then held.  “morphological stasis for about 0.6 m.y.”  Their conclusion:  “The evolution of the Globoconella clade shows both phyletic gradualism and punctuated equilibrium.”  A study of the Globoconella clade of planktonic foraminifera from the late Neogene.  Kennett returned to his own 1981 study of this lineage, which used a single drill site, and this time sampled four Deep Sea Drilling Project sites on a latitudinal transect.  The article uses it because one lineage can change gradually in one place and in sudden bursts in another, so a single site can mislead. ↩ ↩

Pearson, Paul N., and Thomas H. G. Ezard.  “Evolution and speciation in the Eocene planktonic foraminifer Turborotalia.”  Paleobiology 40(1):130–143.  2014.  “We find that for most of the record, contemporaneous specimens form one morphological cluster, which we interpret as an evolving species that shows quasi-continuous but non-directional gradual evolutionary change (anagenesis).  However, in the upper Eocene from ca. 36 to ca. 34 Ma there are two clusters that persistently occupy distinct areas of morphospace, from which we infer that speciation (cladogenesis) must have occurred.”  A study of 10,200 individual tests (the shells) of the Eocene planktonic foraminifer Turborotalia, taken from 51 stratigraphically ordered samples at one site in the tropical North Pacific gyre.  The site was chosen because its oceanography was stable, which removes the water-mass confound that wrecks single-site studies.  The shells were clustered without assigning specimens to species beforehand, which removes the circularity that Hull and Norris complain about.  The article concedes its finding because it is the best-controlled study in this literature, and it found a genuine species split near the end. ↩

Malmgren, Björn A., and James P. Kennett.  “Test size variation in Globigerina bulloides in response to Quaternary palaeoceanographic changes.”  Nature 275:123–124.  1978.  “We report here that changes in mean sizes of Globigerina bulloides can also be used as a palaeoclimatic index in the late Quaternary.”  The same abstract lists coiling direction, pore concentration and test surface ultrastructure as palaeoclimate indicators already in use, and notes that Orbulina universa diameter tracks glacial–interglacial cycles.  A paper in Nature on the size of the shells of Globigerina bulloides, a planktonic foraminifer (a floating single-celled sea animal), through the late Quaternary.  The authors proposed mean shell size as a palaeoclimate index, a measure of past climate.  The article uses it because the same two authors used shell size as a climate indicator three years before they used it as evidence of evolution. ↩

Malmgren, Björn A., and James P. Kennett.  “Phyletic gradualism in a Late Cenozoic planktonic foraminiferal lineage; DSDP Site 284, southwest Pacific.”  Paleobiology 7(2):230–240.  1981.  Sampling:  “The sampling interval is about 0.1 Myr over nearly 8 Myr.”  Verdict:  “Gradual evolution (phyletic gradualism) clearly occurs in all but one measured parameter.  No clear evidence exists for abrupt evolutionary steps (punctuated equilibria) within the bioseries.”  One measured trait moves independently of the environment, and two track it.  The authors' next two sentences say:  “The number of chambers in the final whorl decreases almost linearly, despite known paleoceanographic oscillations within the temperate water mass.  Mean size and apertural shape variations seem to correlate with paleoceanographic change.”  The authors go on to attribute the major changes to a latest Miocene cooling.  A study of one planktonic foraminifer lineage from a single deep-sea drilling site (DSDP Site 284) in the southwest Pacific.  The paper is reference 21 in Hunt 2007, one of the two "most promising examples" that Hunt's shift-model re-analysis failed to support.  The article uses it because Hunt's shift-model re-analysis failed to support this case, and the authors' own sentences say that size variations seem to correlate with ocean change. ↩

Schmidt, Daniela N. et al.  “Abiotic forcing of plankton evolution in the Cenozoic.”  Science 303(5655):207–210.  2004.  “We infer that the observed macroevolutionary increase in test size of planktic foraminifera through the Cenozoic was an adaptive response to intensifying surface-water stratification in low latitudes, which was driven by polar cooling.”  The companion paper is Schmidt, Sabrina Renaud and Bollmann, Paleoceanography 18 (2003), doi 10.1029/2002PA000831.  Its abstract says:  “The amplitude of size variability is directly related to the amplitude of the climatic fluctuations as shown by the stable size-temperature relationship over time,”.  and that the changes were.  “caused by species replacement and intraspecific size variability,”.  which the article reads as habitat tracking and plasticity respectively, in a proportion the authors say depends on the environmental setting and do not quantify.  A study of more than 500 Cenozoic samples and more than one million shells of planktonic foraminifera, tracing how the test size (shell size) of the group changed over time.  The authors read the increase as real evolution driven by the environment, not as mere plasticity, which is a single genotype changing its size with conditions.  The article uses it as the study that makes the environmental case for this trend. ↩

de Vargas, Colomban et al.  “Pleistocene adaptive radiation in Globorotalia truncatulinoides: genetic, morphologic, and environmental evidence.”  Paleobiology 27(1):104–125.  2001.  “The wide geographic distribution of G. truncatulinoides is associated with a latitudinal morphological variability considered as an ecophenotypic variation within a single species.  Here, we present the first molecular, morphological, and ecological evidence that G. truncatulinoides corresponds to a complex of four genetic species adapted to particular hydrographic conditions.”  They add that species from the same region can be told apart by coiling direction, the character Hull and Norris used to find their Miocene form.  The same result holds for the polar workhorse Neogloboquadrina pachyderma, where Kate F. Darling, Michal Kucera, Carol J. Pudsey and Christopher M. Wade report.  “extensive genetic diversity within this morphospecies”.  in PNAS 101 (2004): 7657–7662, doi 10.1073/pnas.0402401101.  A study of the planktonic foraminifer Globorotalia truncatulinoides, whose shell shape varies with latitude, combining genetic, morphological and environmental evidence.  Its variation had been treated as ecophenotypic variation, meaning shape changed by the environment within one species.  The authors found instead a complex of four genetic species adapted to different water conditions.  The article uses it because the variation was not plasticity but hidden species. ↩

Berean Archive.  The Reptile Jaw and the Mammal Ear.  n.d.  A Berean Archive article that assesses, on its own terms, the sequence Gould reaches for, the reptile jaw becoming the mammal ear, and finds that it holds.  It is a class-level transition, so it is the example the article's Gould passage points to. ↩

Eldredge, Niles.  “Misquoted Scientists Respond.”  Creation/Evolution (VI (Fall 1981)).  1981.  “Among other things, the notion of punctuated equilibria accounts for the lack of change seen in most fossil species as they persist through, in some cases, several millions of years,”.  “Our data agree perfectly with the general notion that life has evolved.”  A statement by Niles Eldredge, printed in John R. Cole's piece "Misquoted Scientists Respond."  The article uses it because Eldredge made the same point in the same year, that punctuated equilibrium explains why most fossil species show a lack of change. ↩

Eldredge, Niles, and Stephen Jay Gould.  “Punctuated equilibria: an alternative to phyletic gradualism.”  Models in Paleobiology, 82–115.  Freeman, Cooper, San Francisco.  Edited by T. J. M. Schopf.  1972.  Page 95:  The four tenets are.  “(1) New species arise by the splitting of lineages.  (2) New species develop rapidly.  (3) A small sub-population of the ancestral form gives rise to the new species.  (4) The new species originates in a very small part of the ancestral species' geographic extent—in an isolated area at the periphery of the range.”  The authors also say, on the same page:  “Many breaks in the fossil record are real; they express the way in which evolution occurs, not the fragments of an imperfect record.”  The chapter that sets out punctuated equilibrium, the idea that new species arise rapidly in small populations at the edge of an ancestor's range.  The authors argue that such speciation should leave gaps in the fossil record, gaps that had often been read as evidence of the incompleteness of the record.  The article uses it because the theory was offered to explain stasis, and because the same chapter makes testable predictions. ↩

1. 42a.

   Pages 94–95:  Tracing a fossil species through a local rock column, the authors write,  “should produce no pattern of constant change, but one of oscillation in mean values”.  and, they predict,  “it is likely that the two species will display their greatest difference when the descendant first appears”.  and.  “The allopatric theory predicts that most variation will be found among samples drawn from different geographic areas rather than from different stratigraphic levels in the local rock column.”  Page 114:  Their proposed mechanism of stasis is homeostasis rather than stabilizing selection: species and individuals are systems.  “amazingly well-buffered to resist change and maintain stability in the face of disturbing influences.” ↩

Ricklefs, Robert E.  “Paleontologists Confronting Macroevolution.”  Science 199(4324):58–60.  Jan 6, 1978.  “The punctuated equilibrium model has been widely accepted, not because it has a compelling theoretical basis but because it appears to resolve a dilemma.  Apart from the obvious sampling problems inherent to the observations that stimulated the model, and apart from its intrinsic circularity (one could argue that speciation can occur only when phyletic change is rapid, not vice versa), the model is more ad hoc explanation than theory, and it rests on shaky ground.”  A review published in Science on 6 January 1978 of A. Hallam's edited book Patterns of Evolution as Illustrated by the Fossil Record.  The article uses it because the review calls punctuated equilibrium circular and ad hoc, although the same review affirms that adaptive radiations occur in brief bursts. ↩ ↩

1. 43a.

   “At the family to class levels of taxonomy, most paleontologists agree that adaptive radiations occur in brief bursts, often after the decline of ecologically related groups, and are followed by long periods of evolutionary quiet.”  The review also says:  “Even though Eldredge and Gould may be proven right, their model, and other recent models in paleontology, should not be accorded the status of a major synthesis.” ↩

Valentine, James W., and Douglas H. Erwin.  “Interpreting Great Developmental Experiments: The Fossil Record.”  Development as an Evolutionary Process.  Alan R. Liss, New York.  Edited by Rudolf A. Raff and Elizabeth C. Raff.  1987.  “The populations must remain small (and undetected) and evolve steadily and consistently toward the body plan that comprises the basis of a new phylum (or class).  This is asking a lot.”  The closing question is.  “Why should an opportunistic, tinkering process set on such a course and hold it for so long successfully among so many lineages?”  A chapter beginning on p. 71 of the book and asks what the fossil record shows about the origin of major body plans.  The article uses it as the sharpest objection to rapid origins of body plans, because rapid change would need small populations that stay undetected. ↩ ↩

1. 44a.

   Scope warning: the passage is about the origin of phyla and classes in the Cambrian.  The estimates it opens with are estimates of the rates needed to build a new body plan, not of species-level punctuation rates. ↩

Erwin, Douglas H., and James W. Valentine.  “Hopeful monsters, transposons, and Metazoan radiation.”  Proceedings of the National Academy of Sciences 81:5482–5483.  1984.  “The appearance of many novel morphologies, frequently expressed taxonomically as new phyla, classes, or orders, occurs with such rapidity in evolutionary time that microevolutionary substitutions involving structural genes seem an implausible mechanism.”  A two-page paper in which the authors proposed that RNA viruses passing genetic material between members of a population could quickly create new morphological types.  The article uses it because the same paper gives a statement of the problem and the authors' proposed solution, horizontal transfer by RNA viruses, which went nowhere; quoting the statement while omitting the solution is misquoting them. ↩

Estes, Suzanne, and Stevan J. Arnold.  “Resolving the Paradox of Stasis: Models with Stabilizing Selection Explain Evolutionary Divergence on All Timescales.”  American Naturalist 169(2):227–244.  2007.  As Uyeda et al. 2011 report it, Estes and Arnold.  “found a striking pattern of bounded fluctuations in phenotype, which implies that the expected magnitude of phenotypic change is about the same regardless of whether two samples are separated by 10 generations or 1 million generations,”.  and the model that fit best was.  “displacement of an optimum within an adaptive zone with stable boundaries.”  A paper on the paradox of stasis, which argues that stabilizing selection, the process that holds a trait near an optimum (the trait value that selection favors), explains evolutionary divergence on every timescale.  The article uses it as the only entry in its table with a fitted quantitative model: the authors tested six quantitative-genetic models and found that an optimum moving within bounded limits fits divergence at every timescale. ↩ ↩

Gingerich, Philip D.  “Research on Rates of Evolution.”  Philip D. Gingerich's research page, University of Michigan.  n.d.  Fossil rates run 10^-7 to 10^-3 standard deviations per generation and, rescaled to a one-generation interval, project to 10^-1 to 10⁰, the same range as laboratory selection experiments.  Gingerich's summary:  “For the question — how fast is evolution in the face of selection? — the simple answer is 'fast' and the number to go with this is about one-tenth of a standard deviation per generation.”  A research web page by Philip D. Gingerich summarising his compilations of evolutionary rates, measured in standard deviations of a trait per generation, from laboratory selection experiments, field studies, historical records and fossil sequences.  The underlying compilations are Science 222 (1983): 159–161; American Journal of Science 293A (1993): 453–478; and Genetica 112/113 (2001): 127–144.  The article uses it for the finding that measured rates fall as the time interval grows, and for Gingerich's model in which body size stays within bounds.  Mirror:  Archive.org ↩

1. 47a.

   “Empirically, when rates of evolution are calculated on different time scales, the rates decline systematically the longer the time interval being studied (Gingerich, 1983, 1993, 2001; Figure 1).  This is not an assertion — it is an empirical observation that anyone willing to compile rates over a range of time scales can easily demonstrate for himself or herself.” ↩

2. 47b.

   Gould, Science 226 (1984): 994–995, called the inverse relation.  “a psychological and mathematical artifact”.  Gingerich replied in the same issue (226: 994–996) that.  “arguing that an inverse relationship of rates and intervals exists as an artifact is tantamount to arguing for independence when the opposite has just been demonstrated!”  Others have pressed the point more carefully: Bookstein 1987 (Paleobiology 13: 446–464), Sheets & Mitchell 2001 on spurious self-correlation, and Roopnarine 2003 on random-walk behaviour.  Gingerich answers all three on the page above. ↩

3. 47c.

   “Mammals starting at some average size at the beginning of the Cenozoic can be expected to have diffused and filled the lattice in less than 10⁵ generations … Then they were constrained to evolve within the lattice for the remaining 99 percent of their history.”  His arithmetic: the smallest living mammal and the largest differ by about 100 standard deviations of log body weight, or 10³ units of a tenth of a standard deviation, against roughly 10⁷ generations of Cenozoic history. ↩ ↩

Foote, Michael.  “The Evolution of Morphological Diversity.”  Annual Review of Ecology and Systematics 28:129–152.  1997.  Page 131:  Variance-based measures are preferred because they are insensitive to sample size, whereas range-based measures grow monotonically with it and are therefore biased by preservation.  The standard review of the field, by Michael Foote of the University of Chicago, on how biologists measure morphological diversity, which is the spread of body forms rather than the number of species.  It sets out the machinery: build a character matrix from continuous measurements or discrete homologous characters; compute pairwise dissimilarities and ordinate them by principal coordinates or principal components; measure the spread with a sum of variances, a sum of ranges, a mean pairwise distance or an occupied volume; optionally measure patristic rather than phenetic distance; then bin by time and summarise the trajectory.  The article uses it for that method and for the limits Foote states himself. ↩

1. 48a.

   Page 143:  “Clearly, how fully a morphospace is occupied is potentially very sensitive to the choice of traits (83). …  Even if this does not allow us to say that diatoms are more diverse morphologically than dinosaurs, it does allow us to address whether morphospace was filled more gradually or abruptly in one group versus another.” ↩ ↩

2. 48b.

   Page 130:  “The extent of morphological divergence among taxa generally increases with taxonomic rank (e.g. 16–18, 36, 42, 44, 59, 173, 175), so secular changes in the number of higher taxa (generally, phyla, classes, and orders) provide an obvious index of disparity.”  Foote's references 16–18, 173 and 175 are the Cherry, Case and Wilson body-shape literature, whose point is that morphological divergence is badly decoupled from genetic divergence across distant groups: frogs in different genera differ less in body shape than humans do from chimpanzees.  Findley 1979, Science 204: 434–435, is a published dissent from that analysis. ↩ ↩

3. 48c.

   Page 139:  “Wills et al showed that disparity metrics including range, variance, and distance from basal node on the cladogram all point to comparable disparity in Cambrian and Recent arthropods.” ↩

4. 48d.

   Pages 137–138:  Foote prints both lists, one of clades with pronounced early increases in disparity and one of counterexamples. ↩

5. 48e.

   Page 135:  Cambrian trilobites.  “seem to exhibit pronounced homeomorphy, and they have been notoriously difficult to group into higher taxa (families and superfamilies)”.  whereas Ordovician trilobites are.  “for the most part, easier to sort into morphologically distinct groups”.  and.  “higher taxa in the Ordovician occupy a greater range of morphological space, but with less overlap.”  Nearest-neighbour analysis confirmed greater clustering in the Ordovician.  In trilobites the clumps got sharper later. ↩

6. 48f.

   Pages 138–139:  “the fact that this pattern is soundly grounded in morphological analysis rules out the possibility that it is an artifact of taxonomic practice, such as the erection of higher taxa that diverge early phylogenetically, but only later morphologically.”  He gives five reasons the early-disparity pattern is not an artefact of using discrete characters, of which the third is that in at least one case the pattern.  “was found when characters used to differentiate the higher taxa were omitted.” ↩ ↩ ↩

7. 48g.

   Pages 130–131:  “In a limited analysis of three major clades of Paleozoic marine invertebrates (trilobites, crinoids, and blastozoans), the number of higher taxa (orders and suborders) captured many, but not all, features of the history of morphological disparity revealed by more direct measures.  Some discrepancies are substantial enough to recommend direct morphological analysis.”  Foote's framing of the rank objection is that the relevant issue.  “is not the biological meaning, but rather the information content, of higher taxa.” ↩

8. 48h.

   Page 135:  “Although some preliminary simulations suggest that an increase in morphological clustering could result from simple diffusive evolution and extinction of intermediates, the evolutionary processes responsible for changes in the clustering of morphological distributions need to be explored in greater depth.”  The references behind it are Foote 1990 on trilobite morphospace and Raup & Gould 1974 on stochastic simulation of morphology. ↩

9. 48i.

   Page 134:  On crinoids:  “maximal disparity (measured as mean pairwise phenetic distance) was attained early, and … the morphological extremes reached early in the group's history were scarcely exceeded over the next 200 My, despite the proliferation of hundreds of new genera.” ↩

10. 48j.

    Page 142:  On the Thomas & Reif Skeleton Space, with pairwise combinations of traits:  “nearly all pairwise combinations have been exploited in living and extinct animals, and thus the space seems quite richly occupied.”  And on the Burgess Shale, Thomas & Stewart.  “found that about 90% of designs ultimately used by animals (considered as pairwise combinations of skeletal features) had already been exploited rather early in animal history.”  The underlying primary paper is Thomas, Shearman & Stewart, Science 288 (2000): 1239–1242, titled "Evolutionary exploitation of design options by the first animals with hard skeletons." ↩ ↩

Erwin, Douglas H., and James W. Valentine.  The Cambrian Explosion: The Construction of Animal Biodiversity.  Roberts and Company, Greenwood Village, CO.  2013.  A book about the Cambrian explosion and the construction of animal biodiversity.  The article uses it for the technical word that describes how animal body plans are spread through morphological space, and for its passage on early disparity. ↩

1. 49a.

   Page 340:  “In fact, metazoan morphologies are quite clumped—underdispersed is the technical term—into clades with unique body plans and with significant gaps in architectural style between them, and this pattern continues among classes within phyla and to some extent even among orders within classes.”  Underdispersed is a spatial-statistics term for points more evenly clumped than a random expectation.  Page 216:  A better passage from the same book states the early-disparity claim cleanly.  “the early species of most new major clades were widely distributed across the eventual morphologic range of the group—its morphospace boundaries—and taxa evolving later tended largely to fill in this space rather than extending it by a significant amount.” ↩ ↩

Wills, Matthew A. et al.  “Disparity as an evolutionary index: a comparison of Cambrian and Recent arthropods.”  Paleobiology 20(2):93–130.  1994.  Section Abstract:  “Phenetic methods assessing character-state variability and the amount of morphological attribute space occupied yield similar results for Cambrian and Recent arthropods. …  None of these methods identifies markedly greater disparity among the Cambrian compared with the Recent taxa.”  Hughes, Gerber & Wills 2013 print the pagination as 90–130 in their own reference list, a typo by one of the same authors.  A comparison of the disparity (the spread of body forms, rather than the number of species) of Cambrian arthropods with that of living arthropods.  The authors used four phenetic measures, which do not assume any relationships between the animals, and two cladistic measures, which work within an evolutionary framework.  The article uses it because it tests whether Cambrian arthropods were more varied in body form than living ones, which bears on Gould's Wonderful Life claim. ↩

Hughes, Martin et al.  “Clades reach highest morphological disparity early in their evolution.”  PNAS 110(34):13875–13879.  2013.  Page 13875:  The paper gives the field's central distinction in these words:  “Disparity is conceptually and empirically distinct from diversity.  For example, a relatively small sample of species that differ greatly from one another morphologically (e.g., one species from each order of insects) is likely to be more disparate than a much larger sample of species that are morphologically more homogeneous (e.g., a thousand beetles).”  A meta-analysis of 98 metazoan clades, vertebrate and invertebrate, radiating throughout the Phanerozoic, the span of time with abundant animal fossils.  Each clade's disparity (the spread of body forms) was measured as the sum of variances on successive principal-coordinate axes, all analysed with discrete character matrices for comparability.  Each trajectory was summarised by its scaled centre of gravity, where values below 0.5 mean bottom-heavy, which the paper calls relatively early high disparity.  A supplementary analysis adds 53 living clades.  The article uses it as the largest test of the early-high-disparity claim, and its headline number is a null. ↩

1. 51a.

   Across all 98 clades, the mean centre of gravity is 0.495 and the median 0.501, indistinguishable from the symmetrical null (V = 2,429, P = 0.992).  Taken as a whole, the sample shows no early-high-disparity signal at all. ↩

2. 51b.

   The 63 clades that do not terminate at a mass extinction are significantly bottom-heavy (t = −2.420, P = 0.018).  The 35 clades ending at one of the "big five" are significantly top-heavy (t = 3.901, P < 0.001) because their trajectories were truncated, and the 53 living clades, truncated by the present, are statistically indistinguishable from them (W = 924, P = 0.980).  Of the 63 free-evolving clades, two-thirds were significantly bottom-heavy or early-saturating (χ² = 4.613, P = 0.016).  “there is no clear temporal trend in disparity profile shapes from the Cambrian to the Recent, and early high disparity is the predominant pattern throughout the Phanerozoic”.  The pattern is therefore not a Cambrian phenomenon and not concentrated after mass extinctions. ↩

3. 51c.

   “Over half of our study clades had disparity profiles that were neither significantly top- nor bottom-heavy.” ↩

4. 51d.

   Page 13877:  “major clades are often distinguished from their paraphyletic progenitors because they possess distinct and defining sets of derived characters, or because an extant crown is defined relative to an extinct stem. …  For example, rather than delimiting a clade of Aves from within the paraphyletic nonavian dinosaurs, it would be possible to define a clade of Aves plus some arbitrary 'depth' of theropod dinosaurs.” ↩

5. 51e.

   “Early high patterns inevitably imply an unsampled period of cladogenesis (or the existence of ghost ranges) at the base of the clade, but this either occurs too fast for the available stratigraphic resolution or is not fossilized (11, 50).”  They also note that their own null is conservative against their finding: because clades begin from a single species and extinction is random with respect to the tree, a homogeneous birth–death model with Brownian character evolution predicts top-heaviness, so.  “Our use of 0.5 as a null is slightly simplistic, therefore, but biased against our principal finding”.  “Our results do not allow us to distinguish between ecological and developmental explanations for this phenomenon.” ↩

Van Valen, Leigh.  “Are categories in different phyla comparable?”  Taxon 22(4):333–373.  1973.  A 1973 paper that asks whether the ranks used to sort animals, such as class within one phylum (a major rank of animal classification) and class within another, measure comparable amounts of difference.  Foote's 1997 paper raises the same objection and cites this paper as reference 153.  The article uses it for the objection that higher ranks are human categories, so a claim that gaps lie between phyla may only restate the rule used to name phyla. ↩

Avise, John C., and Glenn C. Johns.  “Proposal for a standardized temporal scheme of biological classification for extant species.”  PNAS 96(13):7358–7363.  1999.  Section Abstract:  “With respect to conveying useful comparative information, current biological classifications are seriously flawed because they fail to (i) standardize criteria for taxonomic ranking and (ii) equilibrate assignments of taxonomic rank across disparate kinds of organisms.”  A 1999 proposal to standardize biological classification by ranking groups by the absolute age of each branching point (the node).  The authors say current classifications are seriously flawed because they do not standardize the criteria for ranks.  The proposed fix is to rank by absolute node age.  A related paper by Avise and Liu appeared in Biological Journal of the Linnean Society 102 (2011): 707–714.  The article cites it for this rank-by-age proposal, set against the claim that phyla are human categories. ↩

Matzke, Nick.  “Meyer's Hopeless Monster, Part II.”  The Panda's Thumb.  Jun 19, 2013.  Section B, parenthesis:  The relevant point sits in a parenthesis in section B; Matzke says the idea that.  “in evolution, first we should see species-level divergence, then genus-level divergence, then family, etc., all the way up to phylum-level divergence”.  is a confusion produced by fusing Linnaean ranks with cladistic thinking, and that ranks are.  “retrospectively applied from the present, looking backwards”.  The objection applies to three arguments in circulation on this side: the "top-down not bottom-up" fossil pattern, the phylum-first argument, and a gap-scaling claim.  It does not destroy them, because disparity is measurable independently of rank, but any argument phrased in Linnaean ranks is answerable in one sentence and should be rephrased.  A blog review of Stephen Meyer's argument, by Nick Matzke, a phylogeneticist (a scientist who reconstructs evolutionary family trees).  The review is self-described as "rough-and-ready."  The article cites him for the taxonomic point, not for his tone, which is contemptuous and invites dismissal.  The article calls his criticism of Meyer a fair one. ↩ ↩

Budd, Graham E., and Sören Jensen.  “A critical reappraisal of the fossil record of the bilaterian phyla.”  Biological Reviews 75(2):253–295.  2000.  A 2000 reappraisal of the fossil record of the bilaterian phyla (the animal groups with a left and a right side).  The article does not summarize its argument.  The article names it as the canonical statement of the stem-group objection, which says the gaps between phyla shrink as more fossils are placed in stem groups (the animals on the branch below a living phylum's last common ancestor). ↩

Springer, Mark S. et al.  “The Adequacy of Morphology for Reconstructing the Early History of Placental Mammals.”  Systematic Biology 56(4):673–684.  2007.  Table 1:  Table 1 lists the superordinal groups that morphology had proposed, including Edentata with Xenarthra and Pholidota.  The bone matrix is the 185 osteological characters of Asher et al. 2003, which the paper calls "the largest currently available morphological data set to include representatives of all extant orders of placental mammals."  A "Points of View" position article in which the team tests whether skeletal data alone can recover the relationships of the placental mammal orders.  The team treated each living order as if it were extinct, which the paper calls pseudoextinction, leaving that order with only its skeletal data as a fossil would have, and compared the resulting trees with trees built from 20 nuclear gene segments for 53 placental taxa.  The article calls it an experiment with a known answer: the skeletal data failed to recover the four-clade tree that DNA supports. ↩ ↩

1. 56a.

   Section Conclusions, page 682:  “We show that as many as 72% of the living placental orders move to a different superordinal group when molecular data are missing and that four of the 18 orders are rendered poly- or paraphyletic.  Superordinal groups are never recovered as monophyletic when treated as pseudoextinct. …  These results suggest that morphological studies of eutherian interordinal relationships have failed to separate homology and homoplasy and have consistently been misled by the latter.”  72% is 13 of 18; the five that stayed put are rodents, lagomorphs, tree shrews, sirenians and proboscideans (p. 679).  The individual placements and support values in the article's table are from Tables 3 and 4. ↩ ↩ ↩

2. 56b.

   Page 682:  “Cases in point include the recovery of Insectivora when either Eulipotyphla or Afrosoricida are pseudoextinct, Volitantia when chiropterans are pseudoextinct, Altungulata when perissodactyls are pseudoextinct, and Edentata when pholidotans are pseudoextinct.” ↩ ↩

3. 56c.

   Page 681:  The 3,818 informative molecular characters were split into 21 partitions of exactly 175 informative characters each, matching the 175-character morphological matrix, and every partition was mapped onto every other partition's best tree.  Molecular onto molecular: 2.9% to 14.9% increase in tree length, mean 6.7%, over 420 comparisons.  Molecular onto the morphological tree: 9.5% to 18.1%, mean 14.2%.  Morphological onto a molecular tree: 15.7% to 25.5%, mean 19.9%.  The authors' conclusion:  “Thus, there is more conflict between the morphological data and any molecular data partition than between all possible pairs of molecular data partitions.  These findings also suggest that … the disagreement between morphological and morphological data is not simply a sample size (i.e., number of characters) problem.” ↩ ↩

4. 56d.

   Pages 673–674, 682:  Their examples of correlated character evolution: Chase et al. 2002 found that canids with small pelvic girdles tend to have large jaw- and neck-muscle attachments; Carrier et al. 2005 found pelvis size negatively correlated with distal limb-bone dimensions; Kangas et al. 2004 found that raising expression of the single protein ectodysplasin in mice changed tooth number, cusp number, cusp shape and position, and crest formation together.  The authors add:  “Numerous characteristics related to these variables are routinely scored in matrices of dental characters.” ↩

5. 56e.

   Pages 682–683:  The authors say that combining the data sets largely overcomes the problem, and where it does not:  “Problems associated with reconstructing relationships among extant placental orders with morphology alone are largely overcome by combining morphological data with molecular data because the phylogenetic signal associated with the relatively large molecular data set overwhelms the conflicting signal from the smaller morphological data set.  However, combined data only mask the problem for extinct taxa where molecular data are still missing.”  Page 682:  A second problem, which the pseudoextinction analyses did not address:  “Another problem associated with real fossils, and one that we did not address in our pseudoextinction analyses, is that fossil skeletons are rarely complete and some or many osteological characters must be coded as missing.”  Page 674:  Their statement of the consequence:  “the failure of morphological data alone to recover a tree compatible with the four-clade interordinal partitioning of placental mammals raises serious doubts about the ability of current morphological cladistic studies to accurately reconstruct relationships for extinct forms.” ↩ ↩

Asher, Robert J. et al.  “Morphology, Paleontology, and Placental Mammal Phylogeny.”  Systematic Biology 57(2):311–317.  2008.  A "Points of View" commentary by three authors who criticize Springer and colleagues' 2007 paper.  They raise four points: the congruence test is judged against a "true" tree, the critics say the paper generalizes from one bone matrix, its emulation of extinction is a heuristic tool, and paleontology (fossils) and neontology (living animals) are both needed.  The article uses it because the critics accept the molecular clades but dispute the scope of the conclusion, and it calls the exchange settled: the critics did not re-run the test and score better. ↩ ↩

1. 57a.

   Page 311:  “The molecular evidence for the four clades of mammals discussed by Springer et al. (2007) is very strong by any measure and is growing at a rapid pace”.  “In terms of the four-clade division of Placentalia, we suspect that they are largely correct.”  Their objection is that condemning a whole data class on a 2007 topology whose root was still disputed was, in their word, "premature." ↩

2. 57b.

   Page 312:  “constraining the morphological data set to support either monophyletic Euarchontoglires or Xenarthra yields trees that are only slightly longer than the shortest tree.  Based on Winning Sites and Templeton tests, and using Asher et al.'s (2003) morphological data set, these topologies cannot be rejected at a significance level of 0.05.”  The critics print the limit themselves in the next sentence:  “a monophyletic Afrotheria, Laurasiatheria, or the four clades simultaneously do conflict with the morphological data.”  So two clades are unresolved rather than rejected, and two genuinely conflict. ↩

3. 57c.

   Page 313:  “there is no objective difference between orders, superorders, or other Linnean ranks, which have been assigned to various mammalian clades quite fluidly.”  On the same page, some superordinal groups (Glires, Paenungulata) were first recognised partly or entirely on morphological evidence, so.  “morphology provides useful information for some mammalian superorders as well.” ↩

4. 57d.

   Page 314:  “The amount and quality of data that can be gathered from living placentals is much richer, including those on organ systems rarely or never preserved in fossils.”  The defence offered for morphology is exactly the data a fossil never has. ↩

Springer, Mark S. et al.  “Morphology and Placental Mammal Phylogeny.”  Systematic Biology 57(3):499–503.  2008.  Page 499:  The opening verdict:  “Asher et al. (2008) make a number of useful points, but as discussed below these do not diminish the main conclusions of our earlier paper.”  A "Points of View" reply by Springer and four colleagues, the same team as the 2007 paper without Burk-Herrick and O'Brien, answering the 2008 commentary of Asher and colleagues.  It says the commentary's points do not diminish the main conclusions of the 2007 paper, and it answers the rank objection by restating the thesis in rank language rather than defending the ranks as natural units.  The article cites it for the reply's concession that morphology has been useful for recognizing most orders and some superordinal groups, and as the other half of an exchange in which neither side disputes the experiment's outcome. ↩ ↩

1. 58a.

   Page 501:  “Asher et al. are correct that most placental orders and some superordinal groups (e.g., Glires and Paenungulata) were first recognized based on morphology.  Morphology clearly provides useful information for some mammalian superorders.”  Page 500:  “There is no doubt that morphology contains phylogenetic signal at the level of interordinal relationships (e.g., Paenungulata, Glires).” ↩

2. 58b.

   Page 501:  The Winning Sites and Templeton result of Asher et al. (p. 312) is not mentioned anywhere in the reply.  The rank objection is answered by restating the thesis in rank language, as follows:  “the majority of placental orders moved to a different superordinal group when molecular data were coded as missing and the four superordinal groups were never recovered as monophyletic”.  rather than by defending the ranks as natural units. ↩ ↩

3. 58c.

   Page 501:  “We performed congruence analyses for 13 of the 20 nuclear genes in Springer et al. (2007) and the postcranial data set of Horovitz (2004) using our previously described methodology.  Morphological data once again emerged as the most incongruent data partition after standardizing partitions so that they contained the same number of informative characters (Fig. 1).”  The marsupial extension, on the same page, compared five nuclear genes against three morphological matrices (Horovitz and Sánchez-Villagra 2003, Luo et al. 2003, Sánchez-Villagra et al. 2007).  “and obtained results that are similar to those for our comparisons involving placental mammals.”  Two of those three matrices are co-authored by Sánchez-Villagra, one of the critics. ↩

4. 58d.

   Page 501:  “The morphological data sets that we have examined for congruence with molecular data are limited to osteological data and it remains possible that soft-tissue morphological characters will prove more compatible with molecular data.  Nevertheless, osteological characters are the primary type of data that are available for fossil mammals.” ↩

Stanhope, Michael J. et al.  “Molecular evidence for multiple origins of Insectivora and for a new order of endemic African insectivore mammals.”  PNAS 95(17):9967–9972.  1998.  Section Discussion:  “Anatomical studies have not yet provided support for such a diverse African clade, and indeed there is not a single morphological synapomorphy that defines Afrotheria.”  Statistics: parsimony tests rejected a monophyletic Insectivora at P < 0.017 on mitochondrial data alone and P < 0.0001 on combined nuclear and mitochondrial data; the Kishino–Hasegawa likelihood test rejected it at P = 0.0091 and P < 0.0001; bootstrap support for Afrotheria reached 100% on combined data.  Their reason for believing the molecules over the anatomy: the alternative would require convergence across mitochondrial ribosomal genes and several unlinked nuclear protein-coding genes of unrelated function, which.  “is unparsimonious in the extreme.”  A 1998 molecular study that sequenced mitochondrial DNA (the small circular genome inside cells) from all six insectivore families and 37 other mammal taxa, with nuclear genes added as a check.  It found that tenrecs and golden moles fall in an African clade with elephants, sea cows, hyraxes, aardvarks and elephant shrews, outside the other insectivores, and it coined the name Afrotheria for that clade.  The article uses it because it named Afrotheria, and because its statistics made the molecular answer more believable than the anatomy. ↩ ↩

1. 59a.

   The features used to define Insectivora,  “such as (i) hindgut simplification, with the correlated absence of cecum, (ii) reduction of the pubic symphysis, and (iii) large maxillary contribution to the orbit, probably represent convergences rather than shared-derived traits.”  Before the molecular work, anatomy had grouped golden moles with tenrecs and moles with shrews, and golden moles had been called.  “spectacularly autapomorphic.” ↩

Sánchez-Villagra, Marcelo R. et al.  “Thoracolumbar vertebral number: the first skeletal synapomorphy for afrotherian mammals.”  Systematics and Biodiversity 5(1):1–7.  2007.  A 2007 paper proposing that the thoracolumbar vertebral number (the count of trunk vertebrae) is the first skeletal synapomorphy (a shared, derived bone feature) for the afrotherian mammals.  Testicondy, permanently internal testes, is the other character usually named; see Werdelin and Nilsonne, Journal of Theoretical Biology 196 (1999): 61–72.  The article uses it for the complication that a morphological shared feature was eventually found, and a vertebral count needs an articulated skeleton to be seen. ↩

Lydekker, Richard.  “Edentata.”  Encyclopædia Britannica 8.  11th edition.  1911.  Section Edentata:  “Only the anteaters and pangolins are absolutely without teeth … but in all the existing representatives of the order teeth are absent from the front of the jaws, while the cheek-teeth are devoid of roots and of enamel …  Practically this is all the definition that can be given to the assemblage, which is possibly an artificial one.”  The same entry already divides Edentata into three suborders, Xenarthra, Tubulidenta and Pholidota, which later became separate orders.  The entry on Edentata, the order of mammals named by Cuvier, in the 1911 edition of the Encyclopædia Britannica, a general encyclopedia, written by Richard Lydekker.  The article uses it as a warning from 1911 that the group might be artificial, which argues against overstating the case. ↩

Collard, Mark, and Bernard Wood.  “How reliable are human phylogenetic hypotheses?”  PNAS 97(9):5003–5006.  2000.  Section Abstract:  “We found that the phylogenetic hypotheses based on the craniodental data were incompatible with the molecular phylogenies for the groups.  Given the robustness of the molecular phylogenies, these results indicate that little confidence can be placed in phylogenies generated solely from higher primate craniodental evidence.”  A 2000 test by Mark Collard and Bernard Wood, who built family trees for living apes and Old World monkeys from skull and tooth measurements and characters, then compared them with the trees that DNA gives for the same groups, whose relationships are well settled.  The trees built from skull and teeth were incompatible with the DNA trees.  The paper's subject is apes and Old World monkeys, and this series defers its implication for hominins to "Lucy and the Pre-Human Hominins"; the article uses it as a method result, because the same kind of skull and tooth characters are used on fossils. ↩

1. 62a.

   Section Results:  “None of the clades supported by 70% or more of the bootstrap samples were compatible with the consensus molecular cladograms.  The hominoid quantitative analysis supported a (Gorilla, Pan, Pongo) clade at 95% …  The papionin quantitative analysis supported a (Cercocebus, Macaca, baboon) clade at 98% …  The analysis of the hominoid qualitative data yielded one clade, which incorrectly linked Gorilla and Pan to the exclusion of the other taxa (92%).”  The qualitative hominoid matrix grouped Homo with Pongo and Pan with Gorilla, the latter supported by four more apparent synapomorphies than the true chimp–human clade. ↩ ↩

2. 62b.

   Section Discussion:  “More problematically, the tests suggest that such analyses can strongly support phylogenetic hypotheses that are misleading. …  the bootstrap-based tests indicate that craniodental data can return impressive levels of statistical support (e.g., 97%) for patterns of phylogenetic relationship that are most likely incorrect.  In other words, cladistic analyses of higher primate craniodental morphology may yield not only 'false-positive' results, but false-positive results that pass, by a substantial margin, the statistical test favored by many researchers.”  Their diagnosis:  “a mixture of the true phylogeny and the phylogenetically misleading effects of convergence, parallelism, reversal, and/or behaviorally induced morphogenesis.” ↩ ↩

Mounce, Ross C. P. et al.  “Sampling diverse characters improves phylogenies: craniodental and postcranial characters of vertebrates often imply different trees.”  Evolution 70(3):666–686.  2016.  Section Abstract:  “a new partition test, based on tree-to-tree distances (as measured by the Robinson Foulds metric) rather than tree length, reveals that relationships inferred from the partitions are significantly different about one time in three, much more often than expected.  Such differences may reflect divergent selective pressures in different body regions, resulting in different localized patterns of homoplasy.”  The same abstract confirms that craniodental characters are significantly more numerous than postcranial ones, and finds no difference in homoplasy levels between the two partitions.  A 2016 study that took 85 vertebrate datasets published from 2000 to 2013 and split each one into craniodental (skull and teeth) characters and postcranial (the rest of the skeleton) characters, to test whether the two halves give the same tree.  The authors' new partition test found that the two halves give significantly different trees about one time in three, far more often than expected, and they suggest that different selective pressures on different body regions may cause it.  The article uses it because teeth are the hardest tissue in the body, and a very large share of named fossil vertebrate species rest on jaws and teeth, so one animal can yield two different answers. ↩ ↩

Sansom, Robert S., and Matthew A. Wills.  “Fossilization causes organisms to appear erroneously primitive by distorting evolutionary trees.”  Scientific Reports 3:2545.  2013.  A 2013 study that removed soft-part characters (the features that usually do not survive fossilisation) from 78 published morphological matrices for over 2,000 living vertebrates and invertebrates, then compared the new trees with the originals.  Of 2,022 nodes (branch points) in the original strict-consensus trees, 49% were recovered after simulated fossilisation against 57% after deleting the same number of characters at random, and not one of 500 random-deletion replicates did as badly (P = 0.002).  The control that proves the mechanism: deleting hard-part characters instead recovered 48%, statistically indistinguishable from random deletion (P = 0.52), and all 50 jackknife re-samplings kept the difference significant.  The article says this belongs in every fossil discussion, whatever else is thought of the rest, because fossils that lose their soft parts are pulled toward the root of the tree and look more primitive than they are. ↩ ↩

1. 64a.

   Figure 1:  The Fig. 1 caption reads:  “61% of 491 taxa shift significantly down phylogenies, from their original position toward the root, vs 39% which shift significantly up.”  Section Abstract:  “Stem-ward slippage, whereby fundamental taphonomic biases cause fossils to be interpreted as erroneously primitive, is therefore a ubiquitous problem for all biologists attempting to infer macroevolutionary rates or sequences.” ↩ ↩ ↩

Sanderson, Michael J., and Michael J. Donoghue.  “Patterns of variation in levels of homoplasy.”  Evolution 43(8):1781–1795.  1989.  Section Abstract:  “Consistency index is highly correlated with the number of taxa included in an analysis, with homoplasy increasing as the number of taxa increases.  This observation is compatible with a simple model of character evolution in which 1) the probability of character-state change increases with the total number of branches in a tree and 2) the number of possible states of a character is limited.”  “Likewise, the level of homoplasy in morphological and molecular data sets does not appear to differ significantly, although there are still too few molecular studies to be confident of this result.”  A statistical study of 60 recent cladistic analyses, which are family trees built from features, across a wide range of plants and animals.  It measured homoplasy, the independent evolution of the same feature in separate lineages, using the consistency index, a score for how well the data fit a tree.  It found that homoplasy rises as the number of taxa (the groups being compared) rises, and that morphological and molecular data sets do not differ significantly, though there were too few molecular studies to be sure.  The article uses it as the honest baseline, because the same abstract that shows homoplasy rising with taxon number also says that morphology and molecules do not differ significantly. ↩ ↩ ↩

Pisani, Davide et al.  “Congruence of morphological and molecular phylogenies.”  Acta Biotheoretica 55(3):269–281.  2007.  “Molecular trees have higher average congruence than morphological trees, but the difference is not significant, and both within- and between-partition incongruence is much lower than expected by chance alone.  Our results suggest that both molecular and morphological trees are, in general, useful approximations of a common underlying phylogeny and thus, when molecules and morphology clash, molecular phylogenies should not be considered more reliable a priori.”  A comparison of 181 published family trees built from molecules with 49 built from morphology, each pair of trees tested for how far they agree.  Morphological trees agreed less well on average, but not significantly less, and both kinds agreed far more often than chance.  The article calls it the broadest survey of the question and gives its finding that morphology is worse but not significantly worse, which is the strongest single statement against the article's own thesis. ↩ ↩ ↩

Sansom, Robert S. et al.  “Non-random decay of chordate characters causes bias in fossil interpretation.”  Nature 463:797–800.  2010.  “loss of chordate characters during decay is non-random:  the more phylogenetically informative are the most labile, whereas plesiomorphic characters are decay resistant.  The taphonomic loss of synapomorphies and relatively higher preservation potential of chordate plesiomorphies will thus result in bias towards wrongly placing fossils on the chordate stem.”  A decay experiment on amphioxus and lamprey larvae, which recorded the order in which their anatomical characters disappeared as the bodies rotted.  The features that best mark a creature's group rotted first, while the ancestral features lasted longest.  The article uses it as the mechanism for a bias in which decay removes the features that identify a group and leaves the ancestral ones, so fossils are placed wrongly on the chordate family tree. ↩

Delsuc, Frédéric et al.  “The evolution of armadillos, anteaters and sloths depicted by nuclear and mitochondrial phylogenies: implications for the status of the enigmatic fossil Eurotamandua.”  Proceedings of the Royal Society B 268:1605–1615.  2001.  Section Abstract:  “The morphological adaptative convergence between New World xenarthrans and Old World pangolins is confirmed. …  [Our molecular dates] question the placement of the purported fossil anteater (Eurotamandua) from the Middle Eocene period of Europe with the Vermilingua and instead suggest the independent and convergent evolution of this enigmatic taxon.”  A DNA study of the living xenarthrans (armadillos, anteaters and sloths), using one nuclear and two mitochondrial genes from eight of the 13 living genera, with molecular dates for the splits between lineages.  It confirmed the armadillos, anteaters and sloths as groups, and it questioned whether the fossil Eurotamandua from the Middle Eocene of Europe belonged with the anteaters, suggesting instead that it evolved its anteater-like form independently.  The article uses it as the first molecular challenge to the fossil's placement, which later anatomy tested. ↩

Rose, Kenneth D. et al.  “Anatomy of Eurotamandua joresi (Mammalia) from Grube Messel, Germany, based on computed tomography, and implications for its relationships.  Part I:  Background and forelimb.”  Palaeobiodiversity and Palaeoenvironments 106:375–408.  2026.  “The segmentation of scapula, humerus, radius, ulna, carpals, metacarpals and phalanges reveals exceptional specializations for scratch digging.  These include numerous features that are strikingly similar to those of palaeanodonts, and to a lesser degree manids, but which contrast sharply with those of the xenarthrans Tamandua and Dasypus.  Other features once thought to ally Eurotamandua with xenarthrans, such as a secondary scapular spine, absence of teeth, and a tubular snout, are not restricted to Xenarthra. …  This new analysis provides compelling evidence that Eurotamandua is related to Palaeanodonta (Laurasiatheria:  Pholidotamorpha) rather than to Xenarthra, thus separating it from xenarthrans at the superordinal level.  Consequently, there is no evidence that Xenarthra ever reached the Old World.”  A study of the forelimb of Eurotamandua joresi, based on computed tomography (X-ray scanning that builds three-dimensional images) of its skeleton, published as Part I of a series.  The fossil was first described by Gerhard Storch in 1981 from the middle Eocene oil shale of Grube Messel.  The forelimb has many features like those of palaeanodonts, an extinct group of mammals, and it contrasts sharply with those of the living xenarthrans, so the authors place the fossil with palaeanodonts rather than xenarthrans.  The article uses it as the decisive anatomical work on the fossil's placement, twenty-five years after the molecular study. ↩

Thewissen, J. G. M. et al.  “Skeletons of terrestrial cetaceans and the relationship of whales to artiodactyls.”  Nature 413:277–281.  2001.  A description of the skeletons of two early Eocene whales of the pakicetid group from Pakistan, from four partial skulls and about 150 loose bones from several individuals.  The bones show that these archaic whales were land mammals, and a cladistic analysis of the skeletal features placed cetaceans closer to artiodactyls (even-toed hoofed mammals) than to mesonychians, without naming a particular artiodactyl family as the closest relative.  It is cited with the report by Philip D. Gingerich, Munir ul Haq, Iyad S. Zalmout, Intizar Hussain Khan and M. Sadiq Malkani, "Origin of Whales from Early Artiodactyls:  Hands and Feet of Eocene Protocetidae from Pakistan" (Science 293 (2001): 2239–2242, doi 10.1126/science.1063902), on the hands and feet of Eocene protocetid whales.  This case belongs to the whale-evolution article in this series, and the article cites the pair only for one narrow point: a morphological hypothesis that survived cladistic scrutiny for decades was wrong, and the tooth characters supporting it were convergent feeding adaptations. ↩

Pilbeam, David.  “New hominoid skull material from the Miocene of Pakistan.”  Nature 295:232–234.  1982.  A short report of new hominoid (ape-family) skull material from the Miocene of Pakistan.  It is cited with Peter Andrews and John E. Cronin's study, "The relationships of Sivapithecus and Ramapithecus and the evolution of the orang-utan" (Nature 297 (1982): 541–546, doi 10.1038/297541a0).  The article uses the pair to show the outline of a case: Ramapithecus was presented from the 1960s into the early 1980s as the earliest member of the human line, on jaw and tooth characters read as hominid, and was reassigned to the orangutan lineage after a partial face and cranium were recovered.  The article states the outline only and makes no claim about the papers' internal arguments. ↩

Drake, Abby Grace, and Christian Peter Klingenberg.  “Large-scale diversification of skull shape in domestic dogs: disparity and modularity.”  The American Naturalist 175(3):289–301.  2010.  “The amount of shape variation among domestic dogs far exceeds that in wild species, and it is comparable to the disparity throughout the Carnivora.  The greatest shape distances between dog breeds clearly surpass the maximum divergence between species in the Carnivora.  Moreover, domestic dogs occupy a range of novel shapes outside the domain of wild carnivorans.”  “The disparity among companion dogs substantially exceeds that of other classes of breeds, suggesting that relaxed functional demands facilitated diversification.”  A geometric morphometric study, which measures shape with coordinates on the skull, of 106 dog breeds, three wild canid species, and a sample spanning the order Carnivora (the meat-eating mammals, including cats, bears and dogs).  It found that the shape variation among domestic dogs is comparable to the whole Carnivora and that the greatest shape differences between dog breeds are larger than the maximum differences between species.  The authors attribute the extreme range partly to relaxed functional demands on companion breeds.  The article uses the caveat to say this shows what morphology can do under intense artificial selection, not what it usually does. ↩ ↩

Strait, David S., and Frederick E. Grine.  “Inferring hominoid and early hominid phylogeny using craniodental characters: the role of fossil taxa.”  Journal of Human Evolution 47(6):399–452.  2004.  Section Abstract:  “Analyses that examined only extant taxa produced most parsimonious cladograms that were inconsistent with the ape molecular tree.  In contrast, analyses that included fossil hominids were consistent with that tree.  This consistency refutes the basis for the hypothesis that craniodental characters are unreliable for reconstructing phylogenetic relationships.”  A cladistic study, which builds family trees from features, of 198 characters of the skull and teeth in living apes and fossil hominids.  The authors ran six analyses, with and without fossil species, and compared each tree with the molecular tree of apes.  Trees built from living species alone disagreed with the molecular tree, and trees that included fossil hominids agreed with it.  The article calls it the strongest counter-argument in its section, because fossils made the morphological tree better. ↩ ↩

Legg, David A. et al.  “Arthropod fossil data increase congruence of morphological and molecular phylogenies.”  Nature Communications 4:2485.  2013.  “We resolve hexapods within Crustacea, with remipedes as their closest extant relatives, and show that the traditionally close relationship between myriapods and hexapods is an artefact of convergent character acquisition during terrestrialisation.  The inclusion of fossil morphology mitigates long-branch artefacts.”  A cladistic study of 753 morphological characters across 309 fossil and living panarthropods, the arthropod group and its close relatives.  Analysed with the fossils included, the morphological tree recovered the molecular answer for arthropods.  The paper also concedes that the long-held grouping of millipedes and centipedes with insects is an artefact of convergence, which the article calls one of the largest homoplasy-driven errors in invertebrate zoology.  The article uses it as the arthropod case, after Strait and Grine, of fossils improving a morphological tree. ↩ ↩ ↩

Li, Ying et al.  “The hearing gene Prestin unites echolocating bats and whales.”  Current Biology 20(2):R55–R56.  2010.  “we here report the surprising finding that the bottlenose dolphin, a toothed whale, is clustered with microbats in the gene tree constructed using protein sequences encoded by the hearing gene Prestin.”  A short report that built a gene tree from the protein sequences of Prestin, the motor protein of the mammalian inner ear, and found that the bottlenose dolphin, a toothed whale, groups with microbats (echolocating bats).  The same issue carries a companion paper by Yang Liu and colleagues, "Convergent sequence evolution between echolocating bats and dolphins", Current Biology 20 (2010): R53–R54, doi 10.1016/j.cub.2009.11.058.  The genome-wide follow-up is Joe Parker and colleagues, "Genome-wide signatures of convergent evolution in echolocating mammals", Nature 502 (2013): 228–231, PMC3836225.  The article uses it as a real case of molecular convergence that misleads: a gene tree puts a dolphin inside the bats, the same kind of failure it attributes to morphology. ↩

Zou, Zhengting, and Jianzhi Zhang.  “No genome-wide protein sequence convergence for echolocation.”  Molecular Biology and Evolution 32(5):1237–1241.  2015.  “genomic signatures of molecular convergence between echolocating lineages are generally no stronger than those between echolocating and comparable nonecholocating lineages. …  Reexamining the previous selection test reveals several flaws and invalidates the asserted evidence for adaptive convergence.”  A second paper by the same authors, "Are convergent and parallel amino acid substitutions in protein evolution more prevalent than neutral expectations?", Molecular Biology and Evolution 32 (2015): 2085–2096, PMC4833076, which used every one-to-one orthologue in 12 fruit-fly and 17 mammal genomes, found:  “the observed amount of molecular convergence is explainable by neutral models incorporating site-specific information of acceptable amino acids.”  A reanalysis of a genome-wide test that had claimed adaptive convergence in echolocating mammals.  It found that convergence between echolocating lineages is no stronger than between echolocating and comparable non-echolocating lineages, and that the earlier test had several flaws.  The article uses it to show that the genome-wide version of the claim did not survive, while Prestin survives as a genuine single-gene case even in the critics' account. ↩

Feigin, Charles Y. et al.  “Genome of the Tasmanian tiger provides insights into the evolution and demography of an extinct marsupial carnivore.”  Nature Ecology & Evolution 2:182–192.  2018.  Figure 3:  From the Fig. 3 caption, verbatim in part:  “we estimate that approximately 34% of the ancestral phenotypic space between the thylacine and red fox has been closed by convergent evolution and 40% has been closed between the thylacine and grey wolf.  These values are significantly greater than would be expected by chance (convratsig C1, P < 0.05, n = 113).  In contrast, the thylacine did not exhibit cranial convergence with other carnivorous marsupials or eutherian felids.”  Their own scaling: the level of convergence is similar to that of Caribbean Anolis lizards, but across more than 160 million years of separation rather than a recent radiation.  A genome sequence of the thylacine (Tasmanian tiger), made from DNA in a preserved pouch young 108 years old, and a comparison of its skull shape with those of other marsupials and placentals.  The genome placed the thylacine at the base of the carnivorous marsupial group, and showed a fall in genetic diversity that began well before humans reached Australia.  The skull method used 30 homologous three-dimensional landmarks on 113 metatherian and eutherian crania (marsupial and placental mammal skulls), Procrustes superimposition (aligning the skulls to a common size and orientation), principal components, and a phylogeny projected into shape space by squared-change parsimony.  The article accepts the skull resemblance to the wolf as real and measured, but says that the claim that the thylacine and wolf are indistinguishable is overstated. ↩

1. 77a.

   Section Abstract:  “In spite of their extraordinary phenotypic convergence, comparative genomic analyses demonstrated that amino acid homoplasies between the thylacine and canids are largely consistent with neutral evolution.  Furthermore, the genes and pathways targeted by positive selection differ markedly between these species.  Together, these findings support models of adaptive convergence driven primarily by cis-regulatory evolution.”  The authors compared observed homoplasious amino acids against a neutral expectation across 11,254 orthologue groups and 691 pairwise node comparisons.  The article calls this result double-edged.  The most extreme phenotypic convergence in mammals leaves essentially no trace in protein-coding sequence, so the molecules stayed honest while the bones lied.  The convergence being regulatory is exactly the mainstream "constrained developmental toolkit" explanation, not evidence against relatedness inference in general. ↩

Chen, Liangbiao et al.  “Evolution of antifreeze glycoprotein gene from a trypsinogen gene in Antarctic notothenioid fish.”  PNAS 94(8):3811–3816.  1997.  “Arctic cod AFGP genes share no sequence identity with the trypsinogen gene”.  A gene study of two fish in different superorders that make near-identical antifreeze glycoproteins from the same simple glycotripeptide repeat.  The notothenioid gene was recruited from a pancreatic trypsinogen, while the Arctic cod genes have different intron-exon organisations and different spacer sequences from it.  The same authors' companion paper, "Convergent evolution of antifreeze glycoproteins in Antarctic notothenioid fish and Arctic cod", PNAS 94 (1997): 3817–3822, PMC20524, covers the two fish as a pair.  The article uses it as the cleanest illustration that convergence can make the finished product identical while the genes still record separate origins. ↩

Chen, Liangbiao et al.  “Convergent evolution of antifreeze glycoproteins in Antarctic notothenioid fish and Arctic cod.”  Proceedings of the National Academy of Sciences 94(8):3817–3822.  1997.  Companion paper comparing purified antifreeze proteins and one gene each from the polar cod and the Antarctic toothfish.  The proteins are near-identical, and the genes differ in structure, spacers and codon spelling, so the authors conclude the two arose separately.  Its assignment of the cod signal sequence was later corrected by Zhuang and colleagues in 2019.  Mirror:  PMC

1. 79a.

   Page 3817:  “Arctic cod AFGP genes share no sequence identity with the trypsinogen gene, indicating trypsinogen is not the progenitor. Second, the AFGP genes of the two fish have different intron–exon organizations and different spacer sequences”. ↩

Conway Morris, Simon.  “Evolution: like any other science it is predictable.”  Philosophical Transactions of the Royal Society B 365:133–145.  2010.  Section Introduction:  “The basis of this view relies on the phenomenon of evolutionary convergence. …  However, much less appreciated is the ubiquity of this convergence, with examples spanning the entire biological hierarchy from molecules to social systems and cognitive processes.”  A review essay by a Cambridge palaeontologist who argues that evolution is remarkably predictable, because convergence, the independent evolution of similar forms, is ubiquitous.  He names frogs and theropod dinosaurs among the examples and argues against the view that evolution is open-ended.  The article uses it to argue that the mainstream explanation of convergence is not a rival to its own argument, since reliably similar pressures would reliably produce similar forms. ↩

1. 80a.

   Section §2(a):  “In fact, it now transpires that there are a series of striking convergences between archosaur and dinosaur morphs that go well beyond the theropod-like forms mentioned above.  Thus, among the archosaurs, we see both carnosaur-like and ankylosaur-like forms, which in a number of cases were previously thought to be dinosaurs.”  He supports the point by quoting Molnar (2008, p. 586). ↩

Sage, Rowan F. et al.  “Photorespiration and the evolution of C4 photosynthesis.”  Annual Review of Plant Biology 63:19–47.  2012.  “C4 photosynthesis is one of the most convergent evolutionary phenomena in the biological world, with at least 66 independent origins.”  A review on the evolution of C4 photosynthesis, which its title links to photorespiration.  Its abstract calls C4 photosynthesis one of the most convergent evolutionary phenomena in the biological world.  The article cites it for the count of at least 66 independent origins. ↩

Figueirido, Borja, and Christine M. Janis.  “The predatory behaviour of the thylacine: Tasmanian tiger or marsupial wolf?”  Biology Letters 7(6):937–940.  2011.  Verbatim in part:  “All the thylacines clustered with the pantherine cats, and they were also classified as ambush predators by CVA, with an average probability of 95 per cent. …  thus bringing into question the degree of ecomorphological convergence between thylacines and wolves.”  Wroe and Milne, Evolution 61 (2007): 1251–1260, found.  “remarkably consistent but differing constraints”.  separating marsupial from placental carnivore skulls across all sizes and diets.  Wroe et al., Proceedings of the Royal Society B 274 (2007): 2819–2828, PMC2288692, whose three-dimensional finite-element models of the thylacine and dingo skulls found.  “considerable similarity, but also informative differences.”  A geometric morphometrics study, which measures shape with coordinates, of the distal humerus (the lower end of the upper arm bone) in carnivorans whose hunting styles are known.  Canonical variates analysis (CVA) classified each animal as an ambush, pounce or pursuit predator from the shape of that bone.  It classified 80.0% of species correctly by leave-one-out testing, which tests each species against a model built without it.  All the thylacines clustered with the pantherine cats, and were classified as ambush predators.  The article uses it as the postcranial evidence that the thylacine was an ambusher, which brings into question how close its convergence with wolves was. ↩

Foote, Michael et al.  “Completeness of the known graptoloid palaeontological record.”  Journal of the Geological Society 176:1038–1055.  2019.  “2800 or more”.  against about 2,200 documented, implying 79% sampled.  The article uses the 75% figure as its clade-level example of sampling, and uses Foote's own wording on what that figure means.  Section Abstract:  The abstract, verbatim:  “We estimate that c. 75% of species in the geographical regions that contribute to the stratigraphic data have been sampled and that, of those species known from more than a single stratigraphic horizon, c. 85% of their original durations, on average, are represented by their global composite stratigraphic ranges.”  A study of how complete the record of graptoloid species is, which the article calls the most rigorous single-clade completeness study in the literature.  It estimates that about 75% of species in the regions that contribute to the stratigraphic data have been sampled, and that about 85% of the original durations of species known from more than one horizon are represented in their global ranges.  The figure is independently corroborated by Sadler and colleagues (2011), who worked from a different version of the same dataset by a completely different method and put total graptoloid diversity at. ↩ ↩

1. 83a.

   “It is important to emphasize that the relevant pool of species consists of those that lived in the regions represented by the preserved sections actually sampled.  No doubt there were many species living only in areas that have no known record …  So, when we estimate that 75% of species have been sampled, we mean 75% of those species that had the potential to be represented in light of where sediment was accumulating in the past, has survived to the present day and has been studied.  Moreover, our conclusions regarding the integrated global record of graptoloids are not meant to imply that local and regional records approach the global record in terms of completeness.”  The article says this sentence dissolves a contradiction in its own earlier notes, and that Foote states it in his own best case. ↩

2. 83b.

   Section Discussion, pages 1050–1051, table 8:  The article's cross-group calibration figures come from this discussion, with the sources Foote and colleagues cite:  marine animal genera at about 0.5 sampling probability per genus per stage (Foote 2003 and Foote and Sepkoski 1999); selected Ordovician groups at 0.6–0.8 per 8 Myr interval (Connolly and Miller 2001); database genera at 78% per stage by Alroy's three-timer method; North American Cenozoic mammal species at 88% sampled and 77% of duration preserved, from Alroy's composite, in their own Table 8; and Cretaceous mammal species at about 10% (Foote et al. 1999).  The authors note that Alroy's data.  “are also high-graded – for example, by eliminating taxa known from just a single faunal list.”  Their explanation of the Cretaceous-versus-Cenozoic mammal difference is:  “The difference is not surprising in light of, for example, the larger ecological footprint of Cenozoic mammal species and the general loss of terrestrial records with increasing geological age.” ↩ ↩ ↩ ↩ ↩ ↩ ↩

Žliobaitė, Indrė, and Mikael Fortelius.  “On calibrating the completometer for the mammalian fossil record.”  Paleobiology 48(1):1–11.  2022.  A paper that tests how complete the mammal fossil record is for the Miocene, about 18 million years, as named fossil species divided by species inferred to have lived, which the article uses to show how the denominator changes the answer.  That inferred figure is built from present-day standing diversity: today's 6,363 living mammal species, multiplied by the number of species turnovers at a mean species duration of about 1 Myr, giving roughly 121,000 expected against 4,505 named, or about 4%; Sensitivity: a 3 Myr duration raises it to about 10%, and a 0.5 Myr duration lowers it to about 2%.  Per order, the figures are Proboscidea 212%, Perissodactyla 140%, Artiodactyla 14.0%, Carnivora 12.6%, Lagomorpha 6.5%, Eulipotyphla 3.4%, Rodentia 2.9%, Primates 1.4%, Chiroptera 0.3%.  The two impossible values above 100% are the paper's own internal evidence that its constant-standing-diversity assumption fails, because modern proboscidean and perissodactyl diversity is a fraction of what it was, and the 4% inherits that. ↩ ↩ ↩

1. 84a.

   “One may wonder why local completeness is expected to be greater than global.  The first, trivial reason is self-selection bias due to methodological constraints.  Advanced statistical approaches require good samples.  Foote and Raup's FreqRat (1996) needs at last four fossiliferous horizons and at least some fossils should be preserved in at least three horizons. …  Evidently, if advanced statistical estimates come only from well-sampled assemblages, they would represent the completeness of those well-sampled assemblages and would not be representative of global completeness on average.”  They also report Foote and Raup's own version of the caveat from 1996.  The article judges this structural claim independent of the exact percentage, and almost certainly right. ↩ ↩

2. 84b.

   The 16% figure is taken from Nyberg and Howell, Basin Research 27 (2015): 1–14, which this paper cites.  “Today, only 16% of Earth's terrestrial land surface is within sedimentary basins; the remainder of the land is in upland areas that will not be represented in the future rock record (Nyberg and Howell 2015), and this 16% does not represent current climatic zones of the Earth uniformly.” ↩

Crampton, James S. et al.  “Second-order sequence stratigraphic controls on the quality of the fossil record at an active margin: New Zealand Eocene to Recent shelf molluscs.”  Palaios 21:86–105.  2006.  “New Zealand has the most complete Cenozoic molluscan fossil record in the Southern Hemisphere…  Overall, about one-third of all once-living Cenozoic species have been sampled, and average per-stage sampling probabilities are between 20% and 50%.”  A study of the shelly marine molluscs of New Zealand from the Eocene to the present, one of the most complete Cenozoic molluscan records in the Southern Hemisphere.  From a synoptic dataset of 5,241 species, it gives an average per-stage sampling probability of 0.20 (95% CI 0.17–0.25) and an overall completeness of 0.32 (95% CI 0.28–0.37), with 0.28 and 0.41 for the Neogene alone.  The article uses it to show that under a third of New Zealand's shelly species are sampled even once, which is why the Californian figure should not be generalised. ↩ ↩

1. 85a.

   The authors' own restriction on this measure:  “The important point is that, in most studies, estimates of completeness have meaning only in the context of a specified subset of life with a particular set of sampling probabilities.  The present study considers only shelly Mollusca; estimates of completeness for planktic foraminifera or annelids will be very different and subject to very different biases.” ↩

Fenton, Isabel S. et al.  “Triton, a new species-level database of Cenozoic planktonic foraminiferal occurrences.”  Scientific Data 8:160.  2021.  Their completeness measure is:  “the fraction of age bins in which a species occurs between its speciation and extinction”.  “Using 1 Ma bins, the median completeness for species is calculated to be 100% – 240 species (62.5%) having a 'complete' fossil record at that resolution.  The mean completeness at this resolution is 87.8%.  At the finer resolution of 0.5 Ma bins, 198 species (51.6%) have 100% completeness, while the mean is 83.4%.”  “The final Triton dataset contains 512,922 non-zero records …  To put our dataset in context, the largest macrofossil dataset at an equivalent taxonomic level are bivalves in the Palaeobiology Database, with 197,606 records for the entire Phanerozoic as of November 2020, and only 79,427 for the Cenozoic.”  Taxon coverage:  “our dataset contains records for 90% of valid species (394 of 438).”  A database paper that compiles species-level occurrences of planktonic foraminifera, single-celled ocean animals with calcareous shells, across the Cenozoic, and measures how completely each species' range is recorded.  Three caveats this article carries: the denominator is a taxonomy, not the world, so the measure inherits exactly Foote's conditional and cannot detect a species that left no record at all; sampling is skewed toward recent intervals and tropical-subtropical latitudes; and the paper's own arithmetic has a small inconsistency, since 240 is 62.5% of 384 species, not of the 394 reported as having records.  The article uses its figures for the one group whose record is unusually complete. ↩ ↩ ↩ ↩

This article, its own computation.  Share of stages occupied by genera in the Paleobiology Database, microfossil and other groups.  n.d.  Section Values at a five-stage threshold:  In descending order, with genus counts: Brachiopoda 0.795 (724), Mammalia 0.789 (329), Trilobita 0.786 (126), Conodonta 0.721 (68), Radiozoa 0.721 (232), Reptilia 0.688 (151), Coccolithophyceae 0.678 (86), Rotaliida 0.670 (57), Aves 0.658 (71), Foraminifera 0.652 (448), Bivalvia 0.639, Bacillariophyceae 0.636 (50), all genera 0.636 (8,291), Mollusca 0.629, Miliolida 0.627 (41), Gastropoda 0.598, Nodosariata 0.595, Echinodermata 0.586, Bryozoa 0.578, Ostracoda 0.520 (202), Angiospermae 0.514.  The ordering is stable at thresholds of three, five and eight stages.  Section Values at an eight-stage threshold:  At eight stages (3,871 genera) the values are mammals 0.820, brachiopods 0.800, radiolaria 0.728, nannoplankton 0.679, conodonts 0.672, foraminifera 0.630, ostracods 0.578, all genera 0.647.  Section Published figures:  Published completeness figures exist for only two of the seven microfossil groups asked about, planktic foraminifera in Fenton et al. (2021) and ostracods in Foote and Sepkoski's Figure 1, which is why the rest were computed here.  The measure inherits every sampling bias documented elsewhere in the article, and it uses a different rank and bin width from the Triton measure it is set against.  A computation done for this article rather than taken from the literature, and nobody has checked it.  It applies Fenton et al.'s measure, the share of stages between a genus's first and last occupied stage in which it actually occurs, at genus rank and International Chronostratigraphic Chart (ICS) stages.  Only occurrences whose entire age bracket falls inside a single stage are used, so a loosely dated collection can neither create nor fill a gap, and the run requires a genus to occupy at least five stages, which leaves 8,291 genera. ↩ ↩ ↩

1. 87a.

   Arithmetic done for this article, introducing no assumption beyond the two measurements it uses.  Boudreau's (1998) 9.8 cm mixed layer divided by Renaudie and Lazarus's (2025) accumulation rates gives 0.33 kyr of deposition blended per sample at 30 cm/kyr, 2.5 kyr at the modern mean of 3.9 cm/kyr, 9.8 kyr at 1 cm/kyr and 19.6 kyr at 0.5.  At 12 to 25 generations a year (Bijma et al. 1990) that is roughly 30,000 to 62,000 generations per sample at the modern mean rate, and 235,000 to 490,000 at the slowest.  Hunt's planktic foraminiferal sequences carry a median of 22 to 48 such samples, so the densest fossil record on Earth watches a lineage through a few dozen windows, each of which has already averaged on the order of a hundred thousand generations into one number.  Nobody in this debate on either side appears to have done this arithmetic, and it constrains both sides. ↩

Foote, Michael, and J. John Jr. Sepkoski.  “Absolute measures of the completeness of the fossil record.”  Nature 398:415–417.  1999.  “These measurements are nonetheless highly correlated, with outliers quite explicable, and we find that completeness is rather high for many animal groups.”  The FreqRat, the measure of preservation probability, is explained in their Methods:  “If f1, f2 and f3 are the frequencies of genera with stratigraphic ranges of one, two and three intervals of geologic time, then the ratio f2²/(f1 f3) estimates the per-interval preservation probability.”  Its stated assumption:  “this measure tacitly assumes homogeneous data, a condition we attempt to satisfy by analysing major taxa independently.”  A study of how complete the fossil record is for animal groups, using two independent measures on two different datasets: the per-interval preservation probability of 16,464 genera resolvable to 103 stratigraphic intervals averaging 5.5 Myr, and the proportion of living families with some fossil record.  The authors found that the two measures are highly correlated, with explicable outliers, and that completeness is rather high for many animal groups.  The article uses it as a fair answer to the crude claim that the fossil record is hopeless, while noting that it says nothing about tetrapods. ↩

1. 88a.

   Figure 1:  The paper prints no table; its per-class values appear only as points in the figure, so any single summary number attributed to it is a reading off a graph.  Values from the published plot run from about 0.06 for polychaetes to about 0.9 for brachiopods, approximate to about ±0.03.  The median across the fifteen classes is around 0.45–0.5, which matches Foote's own later summary of.  “in the neighbourhood of 0.5 per genus per stage.”  The explicable outliers, in the paper's own words, are that cephalopods score low on family recovery because nearly all living cephalopods are soft-bodied coleoids, and that chondrichthyans score high on families and low on genera because families can be extended back on isolated teeth while Palaeozoic genera are whole-body Lagerstätten fossils confined to single intervals. ↩

2. 88b.

   Section Methods:  The fifteen classes analysed, with living-family counts: Anthozoa (122), Asterozoa (53), Bivalvia (103), Brachiopoda (19), Bryozoa (143), Cephalopoda (41), Chondrichthyes (42), Crinoidea (26), Echinoidea (50), Gastropoda (230), Malacostraca (371), Osteichthyes (328), Ostracoda (34), Polychaeta (75), Porifera (107).  There is not a single tetrapod family in the study.  Groups with fewer than 100 resolvable fossil genera or fewer than 10 living families were excluded, and the paper states that the resulting data.  “favour more diverse and better-preserved groups (nemertines, priapulids and many other soft-bodied phyla were excluded, for example).”  The paper also states:  “We did not tabulate the proportion of living genera with a fossil record, as adequate data are not available for most groups.”  Only 68.1% of the 24,190 genera in the source compendium could be resolved to an interval at all. ↩

Valentine, James W.  “How good was the fossil record? Clues from the Californian Pleistocene.”  Paleobiology 15(2):83–94.  1989.  Section Abstract, page 83:  “of which 538 or 77% are known as Pleistocene fossils from the same region.”  “At least 85% of durably skeletonized living species may have been captured in the record”.  This is the second sentence of the abstract.  The hedge is Valentine's, and a paraphrase that hardens "may have been" into "were" misrepresents him.  How the 85% is derived is not audited here, because only the abstract was read.  A review co-authored by Susan Kidwell (Behrensmeyer and colleagues, 2000) independently confirms the 77% figure, and says:  “Valentine (1989) reports that Pleistocene marine faunas in California include 77% of the living mollusk species from the Californian province, with most 'live-only' species being numerically rare and restricted to deeper-water habitats that are not well represented in onshore Pleistocene outcrops.”  Kidwell and Holland's review (2002) adds that 100% of Californian molluscan genera have a local fossil record.  A study of the living shelf shells of California set against their fossils from the Pleistocene, the last two million years.  Its abstract counts 698 living Californian shelf bivalve and gastropod species, of which 538, or 77%, are known as Pleistocene fossils from the same region.  It measures skeletonised species, a different measurement from anything in this article's own family-level run.  The article uses it to ask how many species alive today are already known as fossils, a different question from the completeness of the record, and says the Californian number must not be quoted as the completeness of the marine invertebrate record. ↩

Kidwell, Susan M.  “Preservation of species abundance in marine death assemblages.”  Science 294:1091–1094.  2001.  A reanalysis of 85 molluscan (shellfish) live-dead datasets, which compare the shells that accumulate on the seafloor with the living animals there, from 19 independent studies and 30 study areas, at latitudes from 0° to 54°N and from marsh to middle shelf.  The rank order of species in the dead shells correlates with the living community: an average weighted Spearman r (a 0-to-1 measure of how well two rankings agree) of 0.45 ± 0.03 across all 85, and 0.47 ± 0.04 for the 62 datasets with more than 100 live individuals.  The correlation was 0.54 ± 0.05 for coarse-mesh samples against 0.38 ± 0.06 for fine mesh.  Correlations are positive in 92% of all 85 datasets and 97% of the 62 large ones, and averaged 0.61 ± 0.09 in muddy habitats.  The article uses it to show that dead shells track how common each living species was, and that the older view of strong taphonomic bias was largely an artefact of sieve size. ↩

1. 90a.

   From 24 single-census coarse-mesh datasets with at least 100 dead individuals.  The companion paper, Geology 30 (2002): 803–806, adds the complementary result that death assemblages contain about 25% more species than any single census of the living community after sample-size standardisation, with the excess dominated by numerically rare species.  The article's gloss: richness is time-averaged, abundance is not.  Verbatim, the paper says:  “virtually all (95 ± 5%) species sampled alive also occur dead in the same habitat”.  “most (82 ± 5%) individuals in these death assemblages are from species also censused alive.” ↩

2. 90b.

   The mesh-size finding is the methodological core: fine sieves capture larvae and newly settled juveniles, whose abundance is pulsed and whose shells dissolve within about 100 days, and about 1.5 to 2 mm is the usual lower size limit for taxonomically identifiable fossils.  The article notes the limit of this finding: it is about fidelity, not completeness.  It says a death assemblage faithfully records the local living community, and nothing about how many communities are sampled or about anything without a millimetre-scale mineralised skeleton.  Her conclusion, verbatim:  “paleontologists have by default been focusing on the most reliable portion of the molluscan fossil record.” ↩

Kidwell, Susan M., and Steven M. Holland.  “The quality of the fossil record: implications for evolutionary analyses.”  Annual Review of Ecology and Systematics 33:561–588.  2002.  “∼50% of scleractinian coral, ∼75% of echinoid, and ∼90% of shelled mollusk species found alive today are represented by dead skeletal material in locally accumulating sediments …  100% of Californian molluscan genera have a local fossil record (Valentine 1989), and from a single intertidal zone, Schopf (1978) found that 75% of lightly mineralized and 100% of well-mineralized genera have a known fossil record somewhere in the world.”  “small-bodied forms are most susceptible to loss”.  A review by two paleontologists and geologists of what the field had learned in the previous twenty years about how and why fossils go missing.  It names the traits that predict loss: small body size, aragonitic rather than calcitic microstructure, high organic content and small population size.  The article uses its counts of living species that also appear as dead remains in local sediment. ↩

Valentine, James W. et al.  “Assessing the fidelity of the fossil record by using marine bivalves.”  PNAS 103:6599–6604.  2006.  “For the marine Bivalvia … 308 of the 1,292 living genera and subgenera (herein termed 'taxa') are not recorded as fossils.  These missing taxa are not a random sample of the clade, but instead tend to have small body size, reactive shell structures, commensal or parasitic habit, deep-sea distribution, narrow geographic range, restriction to regions exposing few Neogene marine sediments, or recent date of formal taxonomic description.”  “Of the 106 extant bivalve families, 4 completely lack a fossil record, and these 4 each contain 3 or fewer taxa,”.  The global version of the 1989 Californian study, and the better citation for the global figure, according to the article.  Of the 1,292 living marine bivalve genera and subgenera, 308 are not recorded as fossils.  So 76% of living bivalve genera and subgenera are known as fossils globally.  The missing taxa tend to have small body size, reactive shell structures, a commensal or parasitic habit, deep-sea distribution, narrow ranges, a region that exposes few Neogene marine sediments, or a recent formal description.  At family rank, 4 of the 106 living bivalve families lack a fossil record, each with 3 or fewer taxa, and three of the four are restricted to the deep sea.  The paper's own table totals 26% missing against the 24% in the abstract, because one order is treated separately; the article cites 24%. ↩ ↩

Tyler, Carrie L., and Michał Kowalewski.  “The quality of the fossil record across higher taxa: compositional fidelity of phyla and classes in benthic marine associations.”  PeerJ 11:e15574.  2023.  Their conclusion, verbatim in part:  “dead molluscs appear to be an excellent proxy for all taxa when tracking spatio-temporal patterns and shifts in community structure”.  “the fossil record is likely to be reliable with respect to relative comparisons of composition and diversity.”  A field survey by two paleontologists who counted every macroscopic bottom-dwelling invertebrate, live and dead, at 52 dredge localities in Onslow Bay, off North Carolina, over four field seasons from 2011 to 2013.  It counted 8,798 living individuals and 54,116 dead ones, and 157 living species against 155 dead ones; molluscs are 35.0% of the living individuals and 98.2% of the dead, and arthropods are 33.5% and 0.11%.  The percentages are computed from the authors' table, not stated by the authors, and they put the distortion at a roughly 300-fold relative shift in one step; the survey covers one bay, so it shows the filter in one shallow-sea setting.  The article uses them to show how far the filter distorts a single sample. ↩ ↩ ↩

Behrensmeyer, Anna K. et al.  “Taphonomy and paleobiology.”  Paleobiology 26(sp4):103–147.  2000.  “Other studies of marine and continental biomineralizing groups indicate that the probability of incorporation into a time-averaged death assemblage declines with tissue durability—from ~95% for shelled mollusks to ~75% for echinoids and land mammals and ~50% for marine decapods.”  A review of taphonomy, the study of how remains turn into fossils, by three paleontologists.  It finds that the chance of entering a time-averaged death assemblage (remains from many generations piled into one layer) falls as tissues become less durable, and that land mammals sit at echinoids' figure on that ladder, a statement about bone and tooth durability rather than the much lower count of mammal species that reach a database, which is measured elsewhere.  Durability governs whether remains enter a local death assemblage; a global species count additionally requires that the assemblage sit in a preserved basin, be exposed and have been visited, and the article uses the ladder to set the ceiling on what the record can hold. ↩

1. 94a.

   “Soft-bodied taxa have negligible preservation potential under ordinary environmental conditions…  In marine sands and muds, for example, such taxa constitute 30–100% of species.” ↩

Conway Morris, Simon.  “The Burgess Shale (Middle Cambrian) Fauna.”  Annual Review of Ecology and Systematics 10:327–349.  1979.  “1.  About 16% of the fauna cannot be placed in known phyla…  2.  Some 80% of the fauna is soft-bodied.  The low diversity of normal Cambrian faunas… may be due partly to the scarcity of groups possessing readily fossilizable hard parts.  3.  The pervasive idea that certain groups must have had hard parts to function effectively is negated by the existence of relatives [e.g. Naraoia among trilobites] with fragile cuticles.”  A review of the Burgess Shale fauna, the animals of a Middle Cambrian deposit where soft-bodied animals were preserved.  Its author's own numbered summary gives the shares that matter here, and the article uses his figure of 80% soft-bodied to show how much an ordinary Cambrian shale misses.  Arthropods are 37% of the fauna and 87.5% of it was bottom-dwelling. ↩

Benton, Michael J.  “The Completeness of the Fossil Record.”  Significance 6(3):117–121.  2009.  Page 120:  “The collector curve seen in Figure 3 shows the number of new species on the y-axis against some measure of effort on the x-axis. …  The curve should be sigmoid in shape, represented by a logistic function.”  On new territory he wrote:  “the bird curve and the dinosaur curve are interesting:  they appeared to reach the beginnings of the slow-down about 1980 but have since picked up as search efforts have intensified, especially in new territory such as Australia, South America, India and particularly China.”  Page 121:  He ends the article with:  “the astonishing agreement between evolutionary trees based on molecular (DNA) data and the fossil record, and the fact that the detailed pattern of the evolution of life has not changed much over 200 years of study, suggests that the fossil record is adequate for the purpose we have put it to.  It provides incontrovertible evidence of Darwinian evolution.”  Figure 4, page 121:  The y-axis of his Figure 4 is.  “percentage of discovered species”.  so every curve reaches 100% at 1997 by construction.  It is a history of description effort and cannot settle the question either way.  A popular magazine article in which Benton asks whether the gaps in the fossil record are too big to fill, and defends the record against creationist critics by name.  The article uses his description of the collector's curve, a plot of new species against search effort, and his point that the bird and dinosaur curves picked up again after about 1980. ↩

Benton, Michael J.  “The quality of the fossil record of vertebrates.”  The Adequacy of the Fossil Record, 269–303.  Chichester: Wiley.  Edited by S. K. Donovan and C. R. C. Paul.  1998.  A chapter by Michael Benton of the University of Bristol, in a 1998 book on how adequate the fossil record is.  It summarises the quality of the vertebrate record and tests it with a collector's curve, a plot of new species against search effort, and with family trees built from morphology, which carry no dates, matched against the order of the rocks, applied to 384 cladograms.  The article uses it because creationist writers quote most of its numbers, and it draws on the chapter's distinction between what was fossilized and what once lived. ↩

1. 97a.

   Page 287:  Written in the paragraph after he introduces the collector's-curve technique.  “does not, of course, represent all of the life of the past, since multitudes of species must have come and gone, and yet never have been fossilized.  None the less, those organisms that were fossilized are potentially knowable.”  Page 286:  On the page before, introducing the technique, he describes the collector stopping at the 90% or 95% level because.  “the effort required to find the very last species in an area might well equal the effort expended in finding all the others.”  This is the distinction on which the article's saturation inference fails. ↩ ↩

2. 97b.

   Pages 288–289:  Figure 12.9 plots cumulative new European dinosaur species from 1820 to 1990, flattening near 300, and cumulative new Chinese species from 1920 to 1990, still climbing near 150;  his text gives about 220 and 160 species up to 1990.  He wrote:  “the European collector curve shows a fully developed logistic pattern and the rate of determination of new species of dinosaurs has been levelling off since the 1920s.”  Page 289:  On China he wrote that it had entered an accelerating phase in the 1970s, with.  “only an equivocal hint of a levelling off in the 1990s.”  and closed:  “Until the logistic curve definitely begins to bend over, no one can tell!” ↩

3. 97c.

   Page 287:  “One caveat is that the rate of accretion of new species can be distorted horribly by an enthusiastic splitter (a taxonomist who names species based on small differences).”  His fix is to plot net new species, and he flags the asymmetry it creates:  Pages 287–288:  “the synonymy rate falls off for newer dinosaur names since these have not all gone through the normal processes of reassessment by other workers.”  The article's own recount inherits that bias, which inflates the right-hand end of every curve. ↩

4. 97d.

   Benton's method is stratigraphic-cladistic congruence: cladograms are built from morphology with no age information, so the fit between branching order and stratigraphic order is an independent test.  He ran three metrics on 384 cladograms (174 tetrapod, 147 fish, 63 echinoderm): significant Spearman rank correlation in 50% of tetrapod cladograms against 38% of echinoderm and 25% of fish; RCI at or above 50% in 74%, 78% and 84%; SCI at or above 0.5 in 87%, 95% and 69%.  Pages 294–295:  His summary of the result:  “Benton and Simms (1995) obtained some counter-intuitive results when they showed that continental tetrapods have a fossil record that is as good as, or better than, that of echinoderms…  This surprising result could not have been predicted from observations of the field occurrence of both groups”.  His explanation:  Pages 296–297:  “(1) The relative abundance of specimens at individual fossil localities is no indicator of the completeness of their fossil record on a large scale; this depends on the number of stratigraphic horizons that have yielded fossils and on the packing of those horizons in time.  (2) The fossil record of continental tetrapods has probably been more intensively studied than has that of echinoderms and fishes.”  Page 299:  His strongest and most honest argument in the chapter is an explicit argument from silence.  “Perhaps there have been whole phyla, or even kingdoms, of extraordinary organisms that lived at different times in the past, but which have left no fossil indications…  Such organisms are not impossible, but they are unlikely.  The unlikelihood increases day by day as ever more palaeontological effort fails to turn up any hint of such unknown major groups of macroscopic organisms.” ↩

Bechly, Günter.  “Fossil Discontinuities: Refutation of Darwinism & Confirmation of Intelligent Design (video M7w5QGqcnNs, FOCLOnline, uploaded 11 October 2018, 53 min); Discontinuities in the Fossil Record (video qJI5hTtpO04, Zentrum für BioKomplexität & NaturTeleologie, uploaded 23 September 2019, recorded 29 May 2019, 27 min); Discontinuities in the Fossil Record (video C-yaw80Tq10, IDCentre1, uploaded 18 January 2021, 101 min).”  YouTube.  2018.  At 00:02:00:  At video M7w5QGqcnNs Bechly credits the underlying beachcombing analogy to Paul Nelson.  Three recorded YouTube talks by the paleoentomologist Günter Bechly on fossil discontinuities.  Bechly's professional record is genuine, and the article states it precisely: an OpenAlex author record listing 91 works cited 2,144 times, in Cretaceous Research, Journal of Paleontology, Palaeontology, Zootaxa and others; he erected the new fossil insect order Coxoplectoptera in 2011 and co-erected Permopsocida in Scientific Reports in 2016; he co-edited The Crato Fossil Beds of Brazil for Cambridge University Press; and he has continued to publish taxonomy after joining the Discovery Institute.  The article cites the talks for his accurate description of the method and for the 80% claim that circulates among creationists, which it finds dead as attributed. ↩

1. 98a.

   At 00:03:19–00:03:41:  Video M7w5QGqcnNs, verbatim from the auto-generated English captions, punctuation as generated.  The caption engine's "foot and Sikorsky" is Foote and Sepkoski.  Quotations are from the auto-generated captions, which are word-accurate but not sentence-accurate, and any article quoting them should say so.  “there's a famous paper from 1999 by foot and Sikorsky and they found that we have reached the saturation point in many group of organisms specially in multicellular animals for example in vertebrates we have 80% of vertebrate families are recorded in the fossil record.”  Verdict: DEAD as attributed.  The nearest published match for the number is Michael Denton's 1985 table giving 79.1% of living families of terrestrial vertebrates known as fossils, compiled by Denton from Romer's Vertebrate Paleontology.  Denton's number is a compilation from a taxonomy that is now sixty years old.  At 00:10:27:  The same figure recurs in weaker form in video C-yaw80Tq10.  Section Conclusion:  Bechly's own refereed work cuts against his platform claim on exactly this point: Kohli, Ware and Bechly, How to date a dragonfly: Fossil calibrations for odonates, Palaeontologia Electronica 19.1.1FC (2016), doi 10.26879/576.  In its Conclusion the paper states:  “we lack a crown group fossil record for Epiophlebiidae, Austropetaliidae, Chlorogomphidae, Neopetaliidae, Cordulegastridae, and GSI/Synthemistidae”.  These are six living dragonfly families with no crown-group fossil record at all, in the group Bechly knows better than almost anyone alive.  Every calibration in that paper is stated as a minimum age, which is standard practice because the fossil record is understood to register a clade later than it originated.  The two statements are reconcilable, and the honest form of the completeness argument is Bechly's refereed form, not his platform form.  He passes this site's citability test, but the rule does not transfer: describing fossil dragonflies in Zootaxa does not make a claim about global fossil-record completeness peer-reviewed. ↩

Wang, Steve C., and Peter Dodson.  “Estimating the diversity of dinosaurs.”  PNAS 103(37):13601–13605.  2006.  “Here we estimate the diversity of nonavian dinosaurs at ≈1,850 genera, including those that remain to be discovered.  With 527 genera currently described, at least 71% of dinosaur genera thus remain unknown.”  The discovery rates are:  “Since 1990, an average of 14.8 genera have been described annually, compared with 5.8 genera annually between 1970 and 1989 and 1.1 genera annually between 1824 and 1969.”  A research paper that estimates how many dinosaur genera existed, including those not yet found.  Its method is the abundance-coverage estimator applied to genus occurrence counts, with a point estimate of 1,844 genera; its logistic forecast puts 50% of them discovered at 2037 on the optimistic curve and 2056 on the unweighted one.  The article checks this estimate against the database's own counts.  The abstract, in part: ↩

1. 99a.

   “A genus that left no fossils at all, because of lack of preservation, loss due to subduction or diagenesis, or other factors, can never be discovered, and its presence thus cannot be inferred by the ACE or any similar statistical method based only on abundance counts.  Thus, our estimated diversity provides a lower bound on true original diversity.” ↩ ↩

2. 99b.

   “known diversity is strongly correlated with the number of formations at the subperiod level (R² = 60.3%), however, the correlation of estimated diversity and number of formations is substantially lower (R² = 12.8%).” ↩

3. 99c.

   Comparing their estimate with Russell's larger one, they wrote:  “If both our estimate and Russell's are correct, it would imply that 46% [(3,400–1,844)/3,400] of dinosaur genera have no discoverable fossil record.” ↩

Plotnick, Roy E. et al.  “The fossil record of the sixth extinction.”  Ecology Letters 19:546–553.  2016.  “Of the 773 mammal species in the four most severe categories of the IUCN Red List (those we assume will be going extinct) only 68 (8.8%) were recorded in one or more of the fossil databases.  In comparison, of the 3942 species in the lower concern categories, almost an order of magnitude more (777, or 19.7%) were recorded as fossils.”  “some 38.6% of all assessed cetartiodactyl species are identified in the fossil record, but only 17.9% of threatened species.  For rodents, only 17.3% overall are recorded; this value fell to 4.8% for those that are threatened.  Among bats, a dismal three of the 82 threatened species are recorded.”  Other vertebrates:  “only 3.3% of threatened bird species (26 out of 777) and 1.6% of threatened reptile species (9 out of 551) have a recorded fossil record.”  A research paper that matched 5,515 living mammal species, assessed for extinction risk by the IUCN Red List, against three fossil databases.  Of the 773 species in the four most severe categories, 91.2% have no fossil record, so the article uses it to show that the species now at risk are the ones least likely to be found as fossils. ↩ ↩

1. 100a.

   “The fossil record for genera is markedly better than that at the species level…  Overall, nearly 40% of mammalian genera have been recorded as fossils, with ~58% of all species having their genus described in the record.”  They also flag a historical confound honestly and give both readings:  “The median year of description of mammal species with a fossil record is 1834; about 80% of all mammal species have been described more recently,”.  which may mean the fossil record of living species is essentially fully known, or may reflect.  “a tendency to assign a fossil find to the most familiar member of the recent fauna, a form of 'Pull of the Recent'.” ↩

Peters, Shanan E., and Michael Foote.  “Biodiversity in the Phanerozoic: a reinterpretation.”  Paleobiology 27(4):583–601.  2001.  “It turns out that neither sampling proxy is entirely independent.  Geological formations are more finely divided when fossils are abundant and outcrop area may reflect the availability of suitable ancient habitats for life. …  But analysts must be careful that their sampling proxies are not blunt instruments that 'correct' any meaningful record out of existence.”  Kalmar and Currie, in Paleobiology 36 (2010): 51–60, find the continental record.  “still highly incomplete”.  and.  “reasonably representative”.  at family level, and that its diversification curve is.  “unlikely to be an artifact of this rock bias.”  A reanalysis of marine fossil diversity through time, testing whether its short-term changes follow the amount of sedimentary rock available to sample.  Its rock proxy is the number of sedimentary formations named in the USGS Lexicon of Geologic Names, validated against global outcrop area and percent continental flooding.  First-differenced outcrop area against sampled marine genus diversity gives r = 0.63 to 0.68, p < 0.04.  The article uses it to show what the rock record cannot support, including any reading of the shape of a diversity curve.  Benton, in 2009, gave a balanced reading of this literature: ↩

Diesing, Markus.  “Deep-sea sediments of the global ocean.”  Earth System Science Data 12:3367–3381.  2020.  “48% of the seafloor is calcareous ooze”.  and it is not the same measurement, because the depth cut-off and class definitions differ.  A data paper that maps the seafloor sediment types below 500 m with a machine-learning method, trained on a database of seafloor samples and ocean environment data.  Its Table 4 gives the area of each sediment class for the 329.06 million km² of global ocean below 500 m: calcareous sediment 121.106 million km² (36.80%), clay 101.916 (30.97%), radiolarian ooze 46.778 (14.22%), lithogenous sediment 31.267 (9.50%) and diatom ooze 27.996 (8.51%).  Biogenic ooze totals 59.5%, or 195.9 million km², a third larger than the ~149 million km² of Earth's total land area; the article uses the 59.5% to show how much of the deep-sea floor is biogenic ooze.  The figure of 48% is quoted in oceanography textbooks, ↩

Lloyd, Graeme T. et al.  “Sampling bias and the fossil record of planktonic foraminifera on land and in the deep sea.”  Paleobiology 38(4):569–584.  2012.  “We do this by demonstrating that raw diversity curves derived from the land-based and deep-sea records are strikingly different, but that they each correlate with the intensity of sampling in their respective environments, and thus are ultimately controlled by the structure of the geological record in each setting.”  A study that compares the fossil record of planktonic foraminifera, tiny shelled ocean animals, from land sections with the record from deep-sea drilling.  The article uses it to reconcile the 87.8% completeness figure from the deep-sea Triton database with its own 0.65: the organism did not change, the archive did. ↩

Lloyd, Graeme T. et al.  “Quantifying the deep-sea rock and fossil record bias using coccolithophores.”  Geological Society, London, Special Publications 358:167–177.  2011.  “However, correlations between both richness measures and sampling are extremely strong and a model assuming true diversity to be constant accurately predicts much of observed richness.  We conclude that the deep-sea fossil record, like its land-based counterpart, bears a rock record bias.”  A study that built a species-occurrence database of coccolithophores, microscopic marine plankton with calcite plates, from North Atlantic, Mediterranean and Caribbean drilling, and asked whether observed richness tracks sampling.  The article uses it because its own authors, deep-sea specialists, conclude that the deep-sea record carries a rock-record bias. ↩

Seton, Maria et al.  “A Global Data Set of Present-Day Oceanic Crustal Age and Seafloor Spreading Parameters.”  Geochemistry, Geophysics, Geosystems 21:e2020GC009214.  2020.  “We find that the mean age of oceanic crust is 64.2 Myr.”  A data paper that publishes a global age map of the ocean floor, built from selected magnetic anomaly identifications and a plate-tectonic model.  Its per-basin means, on page 8, are 69.2 million years for the Atlantic and 65.7 for the Pacific, with the oldest crust Jurassic.  These are the mainstream figures under discussion, which the article cites as the mainstream position and not as its own evidence, because its scope excludes absolute dating and works from relative order alone; the structural point survives either timescale, since the exceptional density of the microfossil record is a property of pelagic sediment that has not been destroyed, and that sediment covers only the younger part of the column. ↩

Boudreau, Bernard P.  “Mean mixed depth of sediments: The wherefore and the why.”  Limnology and Oceanography 43(3):524–526.  1998.  “are largely restricted to a narrow surficial zone of marine sediments with a worldwide, environmentally invariant mean of 9.8 cm with a standard deviation of 4.5 cm.”  The independence from sedimentation rate is a separate measurement, by Martin H. Trauth, Michael Sarnthein and Maurice Arnold, Paleoceanography 12, no. 3 (1997): 517–526, doi 10.1029/97PA00722.  The paper is titled.  “Bioturbational mixing depth and carbon flux at the seafloor”.  and its finding is:  “Bulk sediment accumulation rates do not influence the depth of bioturbational mixing under normal pelagic sedimentary conditions.”  So the mixed layer is a fixed thickness of sediment, and the amount of time it blends is set entirely by how fast that sediment accumulated.  A short paper that explains why the mixed depth of seafloor sediment, the top layer stirred by burrowing animals, is so stable: its model of the feedback between food supply, its reactivity and mixing intensity predicts 9.7 cm.  The article uses it for the fixed mixed layer of about 9.8 cm.  The bioturbational effects of deposit-feeding organisms: ↩ ↩

Wise, Kurt, and Thomas J. M. Schopf.  “Was marine faunal diversity in the Pleistocene affected by changes in sea level?”  Paleobiology 7(3):394–399.  1981.  A 1981 paper in a mainstream paleobiology journal, by Kurt Wise with Thomas J. M. Schopf, and the whole of Wise's mainstream peer-reviewed output in the indexed literature.  Beyond it are two Geological Society of America meeting abstracts on the Lance Formation of Wyoming, from 2016 and 2018; GSA abstracts are a secular venue but are not peer-reviewed papers, so they show that he does field work, not that the field work has passed review.  Wise took a B.A. at the University of Chicago and an M.A. and Ph.D. in palaeontology at Harvard under Stephen Jay Gould, a credential printed in the 1995 paper itself, so it is not a creationist self-description passed along second-hand.  The article cites him because he is the only young-earth creationist in its section with a doctorate in the field he writes about. ↩

Cavanaugh, D. P. et al.  “Fossil Equidae: a monobaraminic, stratomorphic series.”  Proceedings of the Fifth International Conference on Creationism, 143–153.  2003.  A conference paper whose title states its claim: the horse family Equidae is one created kind (a baramin) diversifying, which the authors call a monobaraminic, stratomorphic series.  Its stratomorphic character, in their argument, is evidence of change within a kind rather than of macroevolution.  The article cites it for that claim about the horse series, and the paper is the technical version of the argument Wise makes in prose in 1995 and 2002.  See also Reeves (2021) on the status of the method it uses. ↩

Gould, Stephen Jay.  Wonderful Life.  Pages 321–323.  1989.  “is a chordate, a member of our own phylum.”  Gould places Pikaia inside a phylum, and the second claim in Wise's parenthesis, that it sits between two, is not on the pages he cites.  Verdict: misattributed.  Its status has since moved twice in the same direction.  Conway Morris & Caron 2012 made it a stem-group chordate from 114 specimens, and Mussini et al. 2024 reported a dorsal nerve cord, showed the fossil had been interpreted upside down, and placed it in a graded stem series running from vetulicolians to Yunnanozoon to Pikaia to crown chordates.  That is worse for the young-earth reading than 1989 was, not better, because a graded series is precisely what a creationist account says the record does not contain.  Stephen Jay Gould's 1989 book, which Wise gives as the source for his Pikaia line.  The article checks those pages, which say the opposite of Wise's parenthesis, and Gould's conclusion there is that Pikaia. ↩ ↩

Wise, Kurt P.  Faith, Form, and Time.  Broadman & Holman, Nashville.  2002.  “our observation”.  The same book offers this as its observation that a very large percentage of modern species have a fossil record, with no number and no citation, which is the ancestor of the 80% claim in Part 4 and is unusable in that form.  A popular-level book that restates Wise's 1995 framework with markedly more confidence.  One sentence in it does not sit comfortably beside the 1995 concession: stratomorphic intermediate fossils are described as rare and apparently not found at all among the groups with the best fossil record.  The article cites both books and notes the difference, because that is a far stronger statement than "Darwin's third and fourth expectations have been confirmed," and the book gives the reader no hint that its author had written the latter seven years earlier. ↩

Wise, Kurt P., and Donna Richardson.  “What Biostratigraphic Continuity Suggests about Earth History.”  Proceedings of the Ninth International Conference on Creationism, 611–625.  2023.  “Concerning biostratigraphy in general, the global biostratigraphic column seems to be real.”  “This suggests that the order in which the Flood buried fossils bore no relationship to fluctuations in Flood energy.”  Verdict on their own use of the statistic to place Flood boundaries: needs work.  The method is defensible and the robustness checks are real, but a single unvalidated statistic computed from a database with the sampling bias documented elsewhere in the article cannot carry that weight.  The right response is not to quote it but to reproduce it, which the project has the data to do and has not yet done.  A 2023 conference paper by the creationists Kurt P. Wise and Donna Richardson, who do original quantitative work on the same Paleobiology Database the article uses.  For each of 114 globally defined stratigraphic boundaries (100 Phanerozoic stage-to-stage, 10 Proterozoic and 4 Archean), they pulled everything from the unit immediately below and immediately above, reduced both to species-level identifications, and kept only the species occurring on both sides within two degrees of latitude and longitude of each other.  They report robustness checks at 1°, 3° and 4°, and the article uses two of their findings that cut against its own movement. ↩ ↩ ↩

Wise, Kurt P.  “Geology.”  In Six Days: Why 50 Scientists Choose to Believe in Creation.  Master Books, Green Forest, AR.  Edited by John F. Ashton.  n.d.  “if all the evidence in the universe turned against creationism, I would be the first to admit it, but I would still be a creationist because that is what the Word of God seems to indicate.”  The authorized text reads "turned", past tense.  Richard Dawkins quotes it as "turns" in The God Delusion, and essentially every secondary source follows Dawkins.  A chapter by Kurt P. Wise in a book of chapters by 50 scientists, edited by John F. Ashton.  The article cites it for Wise's statement that no evidence would change his creationism, and treats that tension as the strongest thing creationists have to state. ↩

Woodmorappe, John.  “The essential nonexistence of the evolutionary-uniformitarian geologic column: a quantitative assessment.”  Creation Research Society Quarterly 18(1):46–71.  Jun 1981.  “The errors in this work should favor the uniformitarian geologic column. Only a small part of the lower, medial, or upper portions of a geologic period need be present for the area to count as having that geologic period represented there.”  His own statement of method, which is creditable.  Verdict on the measurement: sound.  Verdict on the argument built from it: dead, because the title claims the column does not exist while the numbers measure how many periods are present at a place, which is an outcrop and preservation statistic.  Nobody expects continuous deposition everywhere; erosion and non-deposition are ordinary, and unconformities are mapped and named.  A column that is patchy everywhere but never out of order is what conventional geology predicts.  A 1981 article in a creationist journal by John Woodmorappe, who drew ten world maps, one per Phanerozoic period, on an equal-area projection, working mainly from the compilations of a Soviet team of geologists headed by Alexander Ronov.  His headline results are that two-thirds of the earth's land surface has five or fewer of the ten periods in place, and that only 15–20% has even three periods appearing in "correct" consecutive order.  The article uses it because he deliberately biased his own criteria against himself, which the article calls honest work, and because the title claims more than the numbers measure: they count the periods present at a place, not their order, and he never reports a locality where the order is inverted. ↩ ↩

Woodmorappe, John.  “A diluviological treatise on the stratigraphic separation of fossils.”  Creation Research Society Quarterly 20(3):133–185.  Dec 1983.  “through low antediluvian population (primarily)”.  That explanation has to be quantified or dropped, since a population small enough to leave no fossils is hard to reconcile with a pre-Flood world populous enough to warrant judgement.  The claim that would matter, that index fossils occur in inconsistent relative order between localities, is not what the 479 determinations test, and the mechanism he proposed does not appear to have been developed or tested by anyone since, so it should be described rather than deployed.  A creationist treatise of roughly 37,000 words by John Woodmorappe, published in December 1983, and the ancestor of most creationist writing on fossil order.  He plotted over 9,500 global occurrences of 34 major index fossils on 34 world maps, yielding 479 juxtapositional determinations, in a hand-built global occurrence database two decades before the Paleobiology Database existed.  Verdict on the data collection: sound and genuinely original.  Verdict on the central inference: dead.  That few of 34 index fossils are found together in any 320-kilometre circle is unsurprising and irrelevant, because index fossils are chosen precisely for being stratigraphically restricted, and any one region in any one interval samples a small part of global diversity.  The article uses the paper for the two things that genuinely help: he evaluates and rejects pure chance, selective preservation, differential escape, hydrodynamic selectivity and ecological zonation as explanations of fossil order, which is why he had to propose a mechanism of his own, and he confronts the absence of pre-Pleistocene human fossils directly, explaining it. ↩ ↩ ↩

Wood, Todd C.  “Baraminological analysis places Homo habilis, Homo rudolfensis, and Australopithecus sediba in the human holobaramin.”  Answers Research Journal 3:71–90.  2010.  Section Abstract:  He states the result against interest in his own abstract: the analysis tentatively confirms the common creationist claim that fossil hominids divide into human and non-human categories, but.  “in contrast to many creationist claims”.  H. habilis, H. rudolfensis and.  “most surprisingly”.  A. sediba fall on the human side.  For this series' hominin article, which must not assume the question is settled, the same analysis places A. afarensis in the same group as gorillas and chimpanzees, the furthest possible position from human, reached by a creationist using creationist methods on mainstream character data and published in a young-earth journal.  See also Reeves (2021) on the status of the method.  A 2010 paper in the young-earth journal Answers Research Journal, in which Wood re-examined eight previously published cladistic studies of fossil hominids using statistical baraminology, a creationist method for sorting organisms into created kinds.  The method is to take a published cladistic character matrix, a table of coded anatomical features, compute pairwise baraminic distances, look for blocks of significant positive and negative correlation, and visualise the same matrix by multidimensional scaling.  Wood divided the sample into as many as four groups, and the article uses the paper because it reached a result his own movement did not want. ↩

DeWitt, David A. et al.  “Baraminological analysis places Homo habilis, Homo rudolfensis, and Australopithecus sediba in the human holobaramin: Discussion.”  Answers Research Journal 3:153–158.  Aug 25, 2010.  “are the biased opinions of evolutionists”.  “uncritically accepts an evolutionist's claims of certain anatomical similarities … as hard data, rather than opinion.  He then does a statistical analysis of these opinions as though they were objective data.”  Their strongest point is sound and is not a creationist-only objection, because character coding is the known weak joint of morphological phylogenetics generally, which is one reason the article treats fossil-morphology inferences cautiously throughout.  They also note the small specimen base available for Australopithecus sediba at the time.  Their weakest points are that the paper opens with Psalm 8, 1 Corinthians 15:39 and Acts 17:26 where a morphological argument is needed, which concedes Wood's whole question, and that it quotes mainstream taxonomic authority against Wood's conclusion three paragraphs after declaring evolutionists' anatomical judgements to be biased opinion.  Net verdict: this is a genuine, published, unresolved disagreement inside young-earth creationism about whether a named fossil taxon is human, and it is more credible than a united front.  A reply in the same journal by three young-earth creationists, DeWitt, Habermehl and Menton, to Wood's paper.  They argue that the presumed similarities coded in a cladistic matrix are biased opinion, and that Wood treats them as hard data.  Their conclusion is that Wood's placement of Homo habilis, Homo rudolfensis and Australopithecus sediba in the human holobaramin is a false positive result, most clearly for sediba.  The article uses it for the strongest objection to Wood's method, that a cladistic matrix is built from mainstream opinion, so statistics run on it do not produce data. ↩ ↩

Reeves, Colin R.  “A critical evaluation of statistical baraminology: Part 1 — statistical principles.”  Answers Research Journal 14:261–269.  2021.  Section Abstract:  The methods of Wood and co-workers.  “are based on a shaky understanding of statistical principles, and that their use ought to be abandoned.”  Section Conclusions:  His itemised objections: the assumption that all characters are equal.  “appears to be untested”.  ; the choice of distance metric is rarely discussed against the nature of the variables; the usual metric gives equal weight to the presence and the absence of a character without justification;.  “BDC is not based securely on statistical principles, and the question of true statistical significance of the 'correlations' is sidestepped”.  ; multidimensional scaling has a better pedigree but suffers from metric subjectivity and information loss; and it is unclear the bootstrap is correctly applied.  He is not hostile to the project: he calls the goal.  “worthy”.  and the datasets.  “excellent and painstaking—even heroic”.  , and his stated worry is that misapplied statistics.  “may only give ammunition to unfriendly critics of creationist research.”  Practical consequence for this series: no baraminology result may be reported as a measurement.  The first part of a critique, in the young-earth journal Answers Research Journal, of the statistics behind statistical baraminology, the method Wood used.  Reeves is not hostile to the project, and he also notes that evolutionists misapply statistics too.  The article uses it because the method was criticised from inside creationism, so no baraminology result may be reported as a measurement. ↩ ↩

Wood, Todd C.  “The truth about evolution.”  toddcwood.blogspot.com.  Sep 30, 2009.  “Evolution is not a theory in crisis.  It is not teetering on the verge of collapse.  It has not failed as a scientific explanation.  There is evidence for evolution, gobs and gobs of it.”  He goes on to say that people who say evolution is a failure are.  “either unacquainted with the inner workings of science or unacquainted with the evidence.”  A blog post by Todd C. Wood, dated 30 September 2009, in which he says that evolution is strongly evidenced but that his rejection of it is a faith choice grounded in what he takes Scripture to reveal.  The article uses the post because he separates the strength of a theory from its truth, and says why he does so. ↩

Prothero, Donald R.  “Stephen Meyer's fumbling bumbling amateur Cambrian follies.”  Skeptic 18(4).  2013.  “completely ignores the existence of the first two stages of the Cambrian (nowhere are they even mentioned in the book, or the index)”.  Verdict: overstated, see ¹²⁰.  Prothero also records a counter-charge that a Discovery Institute reply quote-mines Erwin and Valentine's book, and that Erwin, asked directly, "found it laughable."  The article does not repeat it as fact, because it is Prothero reporting a private conversation that has not been checked with either party.  A review of Meyer's book in Skeptic, posted at skepticblog.org on 28 August 2013.  The article cites it for the charge that Meyer ignores the early Cambrian stages altogether, a charge it finds overstated. ↩

Meyer, Stephen C.  Darwin's Doubt: The Explosive Origin of Animal Life and the Case for Intelligent Design.  HarperOne.  Paperback edition.  2014.  Section Full text of this edition:  Word counts over the main text, endnotes and bibliography: "small shelly" 8 occurrences, all in endnotes and the bibliography, none in the main text or in any figure; "Nemakit" 0; "Tommotian" 1, in the 2014 epilogue only; "Manykaian" 1, in endnote 39; "Atdabanian" 1, in the epilogue only; Cloudina 0.  Endnote 39 describes the small shelly fossils accurately, with their stage and dates, citing Cooper and Fortey's "phylogenetic fuse" paper.  A book arguing that the Cambrian explosion, the sudden appearance of many animal body plans in the fossil record, is best explained by intelligent design.  The 2014 paperback contains a long epilogue answering Charles Marshall's review, which was published on 20 September 2013, three months after the hardcover, and the 2013 hardcover has no such epilogue.  The article uses it because two critics charged that it ignores the early Cambrian small shelly fossils, and the book's own text is the test of that charge. ↩ ↩ ↩ ↩

1. 120a.

   Chapter 3:  Meyer introduces Marshall as summarising the artefact explanations of the Cambrian explosion and gives a quotation in which two sentences are followed by an ellipsis and then a third.  Meyer's endnote 13 reads: "Marshall, 'Explaining the Cambrian "Explosion" of Animals,' 357, 372."  The source is Charles R. Marshall, Annual Review of Earth and Planetary Sciences 34 (2006): 355–384.  The two page numbers are fifteen pages apart in a thirty-page review.  The ellipsis is present and both page numbers are cited, so nothing is concealed from a reader who checks the endnote; this is not fabrication and should not be called that.  But the passage reads as continuous prose from a single argument.  Verdict: context-stripped, disclosed only in an endnote.  Whether the two pages make the same argument or opposite ones is the question that actually matters, and it is unanswered here. ↩

Marshall, Charles R.  “When prior belief trumps scholarship.”  Science 341(6152):1344.  Sep 20, 2013.  “new phyla were not made by new genes but largely emerged through the rewiring of the gene regulatory networks (GRNs) of already existing genes”.  The third charge is the one the article takes no side on.  Meyer's counter, that rewiring presupposes rather than explains the information in the pre-existing networks, is a real objection, and Marshall's reply that early networks were less encumbered is unfalsifiable as stated.  It is also mostly a molecular argument rather than a fossil one, and therefore outside this series' scope.  The fourth charge is.  “a systematic failure of scholarship”.  grounded in the chimera-quote example.  A review of Stephen Meyer's Darwin's Doubt by Charles R. Marshall, Director of the Museum of Paleontology and Professor of Integrative Biology at UC Berkeley.  The article uses it for Marshall's four charges against the book, in his order: that the approach is a god-of-the-gaps argument (philosophical, out of scope here, and answered at length in Meyer's 2014 epilogue); that Meyer omits the small shelly fossils and misunderstands molecular phylogenetics; that the information argument targets the wrong thing, because. ↩ ↩