Brain Size and the Order of the Fossils

"Even with the drawbacks, there is definitely a trend in encephalization over radiometric time" — Todd Wood, young-earth biologist, 2016¹

Summary

The fossil skulls of the human family get bigger-brained as they get younger.  That pattern is real.  It does not come from how the fossils were named, and young-earth writers have accepted it without yet explaining it.  Two small-brained species from late in the record, Homo naledi and Homo floresiensis, show that brains did not grow in every population.  Large brains early are another matter:  one skull, redated in 2026, is the only real candidate.

1. The trend is large and steady. Andrew Du and colleagues compiled 94 dated fossil skulls from 3.2 to 0.1 million years ago on the standard scale.² Run without any species names, brain volume rises by a factor of about 1.5 per million years, and date alone explains 77 percent of the variation in log brain volume. Of the 3,428 pairs of skulls whose date ranges do not overlap, the younger skull has the larger brain in 91 percent.³ Lucy's species averaged 445 ml;⁴ early Homo sapiens 300,000 years ago already had brains of 1,375 to 1,491 ml, inside the range of living people.⁵
2. It is not an artifact of sorting. The charge has a real basis: brain size was one of the working criteria for putting a fossil in Homo,⁶ and one young-earth paper's line between human and ape turns out to be a pure brain-size cut. But the order by date alone, with no labels, gives the same answer. It survives dropping damaged skulls, dropping doubtful old estimates, keeping only East African sites dated by volcanic ash, and putting skull KNM-ER 1470 back at the 2.9 million years once claimed for it.⁷ What depends on the labels is how the trend is divided: between half and nine tenths of it happens inside named species, depending on how many species one counts.²⁸
3. It is a rising ceiling, not a rising floor. None of the 12 skulls older than 2.0 million years exceeds 550 cc, and on the dates Du's table uses, none older than 1.0 million years exceeds about 1,060. Small brains go on late: Paranthropus to 1.4 million years, late Javanese Homo erectus at 900 to 950 cc, naledi and the hobbit.⁹ One skull now tests the ceiling. Yunxian 2 in China, about 1,143 cc,¹⁰ was redated in February 2026 from about one million years to about 1.77 million.¹¹ At that date it sits 3.5 standard deviations above the Homo trend, the only large brain early. It lowers the overall pattern only from 91 to 90.5 percent of pairs, but young-earth writer Peter Line is right that it is the kind of find the trend should not contain.¹²¹³
4. About half of it tracks body size. On Todd Wood's own numbers, body growth predicts about 70 percent of the rise from Lucy's species to Homo erectus and almost none of the rise from erectus to modern humans. An independent table of 254 dated body masses gives the same half overall, and about a fifth inside Homo.¹⁴ Body masses for fossils are uncertain by about a quarter, which moves these shares widely.¹¹⁵
5. Young-earth writers accept the order. Wood calls the trend real and asks what causes it.¹ In 2010 he wrote that "Any model of post-Babel distribution must account for the appearance and disappearance of these human species in stratigraphic order", and proposed that the dispersal from Babel was led by H. habilis and H. rudolfensis, among the smallest-brained members of his human group.¹⁶ Marcus Ross, Peter Brummel and Wood proposed in 2023 that the first generations after the Flood were highly variable, and admitted that their preferred model "gives no account for either the late appearance of H. sapiens in the fossil record" or its takeover.¹⁷ No young-earth writer found here has published a mechanism that makes brains grow in post-Flood time, or priced one.
6. Both young-earth readings carry a price. If the human kind's brains grew, the average had to more than double within the 292 to 1,172 years between the Flood and Abraham, a sustained response as fast as laboratory mice bred for size, or several times faster, in every population at once. If large-brained people were there all along, they left almost no skulls in the early layers: on the dates Du's table uses, 25 Homo skulls older than one million years and none above 1,062 cc, so that even counting Yunxian 2, large brains can have been at most about one in seven of the early dead.¹⁸
7. The small, late brains cut both ways. The hobbit's 426 cc and naledi's 465 to 560 cc sit 5 to 7.5 standard deviations below the Homo trend for their dates.¹⁹ For the mainstream they show that brain size could stall or shrink on side branches, so the trend is a pattern and not a law. For young-earth writers who call them human, they show human brains as small as an australopith's late in post-Flood history, which supports "variation, not progress" and makes brain size useless for sorting humans from apes. They do nothing to explain why large brains are so rare early.

Verdict:  the rise of brain size with date is one of the clearest patterns in the hominin record, and it is not manufactured by naming or by one disputed date.  The young-earth side has conceded the pattern and offered models of dispersal and early variability that do not yet predict it.  The small, late brains are a real complication for any story of steady growth, but they bend the floor of the trend.  Only Yunxian 2, if its new date holds, bends the ceiling, and one skull does not undo 130.

This article tests the pattern across all the fossils.  Each species has its own article, linked from Hominin Fossils;  how young-earth writers draw the line between ape and human is argued in Ape or Human? How Creationists Sort the Fossils.

Part 1: The data

What is measured.  Brains do not fossilise, but the braincase does.  Its inside volume, the endocranial volume or "cranial capacity", is measured by filling the skull with seed or water, or from a CT scan, and it tracks brain size closely.  Du's team call it "a reliable proxy for brain size".²  It is not exact.  Of the 94 skulls below, 54 have two or more published volumes, and the estimates for one skull differ by a median of 8 percent and at most 30 percent.³  This article writes cc and ml interchangeably;  they are the same unit.

The compilation.  Du and seven colleagues, including Bernard Wood of George Washington University, gathered every published volume for skulls "whose geological age ranges encompass dates anywhere between 3.2 and 0.5 Ma (i.e. beginning with the earliest uncontroversial hominin taxon, Australopithecus, and ending when the ECVs of fossil crania begin to overlap with the range of modern humans)".²  Their table gives each skull's volumes, the source of each, and a minimum and maximum date with the paper the date comes from.²  The table stops where fossil brains reach the modern range, so the later skulls here come from a second source:  Simon Neubauer, Jean-Jacques Hublin and Philipp Gunz measured 20 fossil Homo sapiens, 8 Neanderthals, two Middle Pleistocene skulls and three late Javanese H. erectus from CT scans in one laboratory by one method, with a published age for each.⁵  The two naledi composite skulls and the hobbit, LB1, are added from their own papers.  All ages are on the standard radiometric scale, as published.  What a young-earth chronology does with them is Part 4.

Living humans.  Neubauer's comparison sample of 89 "cranially diverse" present-day people, measured the same way as the fossils, averages 1,328 ml, with a standard deviation of 164 and a range of 1,082 to 1,832.⁵  The largest sample with a published mean is Beals, Smith and Dodd's 1984 survey.  Liberato De Caro, reviewing the figures in 2024, reports that it "analyzed data on 5,288 crania from 122 different ethnic groups and reported an endocranial volume of 1,349 ± 78cc."²⁰  The Beals paper itself was not read for this article, but the familiar "about 1,350" for living people has a real sample behind it.  Other figures differ because they measure other things:  brain tissue measured on MRI is smaller than the cavity that holds it, and a weighed brain is a mass, not a volume.²¹

The first thirteen rows are Du's table, grouped by Du's narrower ("split") names;  the averages are geometric means, the usual choice for sizes.  The next two rows and the four sapiens and Neanderthal rows are Neubauer's;  the naledi and hobbit rows are from Lee Berger's team, Daisuke Kubo's team and the dating papers.^22232425  Two cautions go with the table.  The P. robustus row includes Robert Broom's 1952 estimates of 800 and 750 cc for SK 46 and SK 48, far above every later figure for the species;² Part 2 shows the result does not depend on them.  And the late rows mix laboratories:  Neubauer's volumes are virtual casts with a 23 ml correction added,⁵ Du's are whatever each original author published.

The same 94 skulls, sorted only by date, with no species names at all:

Each skull sits in the bin of its date's midpoint.³  Read down the "Largest" column and the ceiling rises in almost every step.  Read down "Smallest" and it rises too, until the last row, where the hobbit and naledi drop it back to the level of 3 million years ago.

Part 2: A real trend, or an artifact of sorting?

The charge, and its basis.  It comes in two forms.  The first is that species are named partly by brain size, so a row of species set out in order will show rising brains by construction.  That has a basis.  Bernard Wood and Mark Collard wrote in 1999 that "in practice fossil hominin species are assigned to Homo on the basis of one or more out of four criteria.  The first is an absolute brain size of 600 cm3", and that "none of these criteria is satisfactory."⁶  The second form is that dates are picked to fit the expected order.  That has a basis too.  When skull KNM-ER 1470 was first placed at about 2.9 million years, Marvin Lubenow wrote, "The theory of human evolution did not allow for a skull so modern in morphology to be that old."²⁶  Of the dates thrown out as impossible in that dispute, he asked, "How did they know?  The answer is that the associated fossils told them so."²⁶  The full story, including where Lubenow is right, is told in Homo habilis and Homo rudolfensis.

Both charges can be tested with the numbers.

Test 1:  drop the names.  Everything in this test uses only each skull's volume and its date range.  Date alone explains 77 percent of the variation in log brain volume across Du's 94 skulls, and the rank correlation between age and volume is −0.89.  The cleanest check needs no exact dates at all.  Take every pair of skulls whose date ranges do not overlap, so that one is certainly younger, and ask which has the bigger brain.  There are 3,428 such pairs.  In 3,129 of them, 91 percent, the younger skull is the bigger-brained.³  No species label enters that count.

Test 2:  take out what is doubtful.  Dropping the ten skulls Du's team flagged as incomplete or distorted leaves 92 percent.  Dropping the three old robustus estimates of 1946 and 1952 leaves 91 percent.  Keeping only the 38 East African skulls, whose dates are cited to argon dates on volcanic ash and to the ash layers' chemistry, leaves 84 percent, with date explaining 58 percent of the variation.  And restoring 1470 to 2.8 to 3.0 million years, the date Lubenow says evolutionary theory could not allow, leaves 90.7 percent.  That one change does raise the largest brain older than 2.5 million years from 550 to 756 cc, so it would have made a real exception.  But the best-documented case of a date steered by expectation moves the overall pattern by less than one point.⁷

Test 3:  a second compilation.  Jeremy DeSilva and colleagues compiled their own table of 985 dated volumes from the literature in 2021, to ask a different question.²⁷  After removing repeated rows, the ape Rudapithecus, and the Holocene skulls, it holds 283 fossil hominins.  Brain volume rises in it by a factor of 1.48 per million years, against 1.50 in Du's table, and date explains 75 percent of the variation.²⁸

Test 4:  one pile of rock.  Inside a single sequence of layers, which is older does not depend on any dating method, only on which layer lies on top.  At Olduvai Gorge, the lowest skulls, from Bed I, are 509 to 685 cc;  OH 9, thought to come from Bed II above them, is 1,062;  but OH 12, from Bed IV higher still, is 691.²⁹  In the Turkana Basin, the layers below about 2.3 million years have yielded only Paranthropus aethiopicus skulls, at 410 and 427 cc;  from the level of the KBS Tuff upward, Homo skulls of 738 to 909 cc lie beside Paranthropus skulls of 430 to 503.³⁰  Ian McDougall dated seven ash layers through the Koobi Fora part of that basin and reported that the "Results are consistent with the stratigraphic sequence".³¹  These single-sequence samples are small.  They point the same way as the whole table, and they show that the big-brained skulls sit above the small-brained ones in the same rock, but they would not establish the trend on their own.

What the names do decide.  Names decide how the trend is divided, not whether it exists.  Du's team split the rise into change inside lineages and change from species appearing and disappearing.  Inside lineages they found "64% or 88% depending on whether one uses a more or less speciose taxonomy", supplemented by "the appearance in the fossil record of larger-brained Homo species and the subsequent disappearance of smaller-brained Australopithecus and Paranthropus taxa."²  A simpler measure here, the slope inside named species against the slope overall, gives 53 percent with Du's narrow names and 69 percent with the broad ones.⁸  Most of the change inside a species is inside Homo erectus, from 546 cc at Dmanisi to 1,300 at Zhoukoudian.  Susan Antón found in 2003 "a significant but gradual trend for increasing brain size through time in Asian H. erectus, at the rate of about 160 cc per million years."²⁹

Is it only body size?  Bigger animals have bigger brains, so a fair question is whether hominins simply grew.  Todd Wood asked exactly that:  "Is the trend merely an allometry of increasing body size?"¹  On Wood's own taxon means and his own primate scaling line, the answer is "partly".  From Australopithecus afarensis to Homo erectus, average brain volume rose 2.29 times and body mass 2.08 times;  body growth alone predicts a brain 1.77 times larger, about 70 percent of the rise.  From erectus to sapiens, brain volume rose 1.52 times while body mass did not change, so body size explains almost none of it.  Over the whole span, about half.¹⁵  A second check, independent of Wood's averages, agrees.  Manuel Will, Adrián Pablos and Jay Stock compiled every published body-mass estimate for fossil hominins with its date, choosing to work by date because "time is a more reliable variable than taxonomy."¹⁴  On their 173 estimates dated within the span of Du's table, body mass rises by a factor of 1.27 per million years.  That predicts 46 percent of the brain slope across all hominins, and 21 percent inside Homo.¹⁵  These shares are soft.  Wood himself notes that body-mass estimates carry intervals of at least 12 kg, "about 25% of their estimate",¹ and moving one taxon's mass by a quarter swings the first share between 42 and 90 percent.  David Lordkipanidze's team found the same at the start of Homo:  the small Dmanisi Skull 5 has an encephalization quotient "within the range of variation for Australopithecus."³²

A young-earth sort that is a brain-size sort.  The circularity charge applies on the young-earth side too.  Jean O'Micks clustered fourteen hominin taxa by their mean brain volume and mean body mass in 2017.  H. naledi and H. habilis fell with the australopiths, and he concluded:  "This serves as further evidence supporting the idea that H. naledi should be classified as an ape and is non-human."³³  He did not scale the two measures, and in his units brain volume varies eighteen times as much as body mass.  Re-running his clustering on brain volume alone gives exactly the same three groups.³⁴  So his line between human and ape is a brain-size line, and it cannot then serve as independent evidence of a gap in brain size.  O'Micks did state the limit himself:  "just because a hominin species has a small cranial capacity does not automatically mean that it is an ape or less intelligent, the converse is also true, namely that it cannot be assumed that if a species has a small cranial capacity, then it is not an ape."³³

What survives.  The order by date survives every test that can be run on it:  without names, without doubtful skulls, in one region, in a second compilation, and with the most disputed date put back.  What does not survive is any claim that brains grew at the same rate everywhere, or in step with body size, or inside every lineage.

Part 3: What young-earth writers say

The challenge.  In 2006 the biologist Nick Matzke plotted 602 published skull measurements against date on the Panda's Thumb blog and wrote:  "Creationists should be embarrassed and ashamed about the huge mass of evidence they ignore every time they fail to mention the stunning, overwhelming, transitional, gradual, nature of the hundreds of ancient fossil skulls".³⁵  Todd Wood, a young-earth biologist, wrote ten years later that the chart "has become a recurring feature of anticreationist literature" and that it "does strikingly illustrate a lack of discontinuity between what we creationists would recognize as human and nonhuman.  As such, it still has value and warrants a response."¹

The common frame.  Young-earth writers agree on where these fossils sit.  Marcus Ross, Peter Brummel and Wood list the reasons, among them that hominins "appear only at the very top of the fossil record in the most recent sedimentary deposits", and conclude:  "These general considerations strongly suggest that all hominin remains are post-Flood."¹⁷  So on their view the order in this article is an order inside post-Flood history, squeezed from three million radiometric years into a few centuries.  Whether radioactive decay can be squeezed that far is argued in the species articles;  this part asks only what young-earth writers make of brain size inside that frame.  Five answers are in print.

1.  Brain size does not decide who is human.  Marvin Lubenow, whose Bones of Contention is the most widely read young-earth book on human fossils, gives the modern range as "about 700 cc all the way up to about 2200 cc" and says the spread "has virtually nothing to do with intelligence".³⁶  Christopher Rupe and John Sanford:  "Brain size is not what defines humanness."³⁷  By 2026 Rupe had dropped "small cranial capacity" from the traits that mark an australopith, "since they have been shown to occur in human fossils."³⁸  Peter Line, who writes on human fossils for Creation Ministries International, counts 48 Homo erectus skulls from 546 to 1,260 cc and concludes "there was a huge natural variation in cranial capacity of these humans".³⁹  These answer whether a small-brained fossil can be human.  They say nothing about why small brains come first.

2.  The fossils are not in order.  Lubenow's own test was to "place all of the relevant fossil material on a time chart according to the evolutionist dates", and he found "a morphological continuum, both horizontally across species and vertically over time.  The horizontal continuum shows that anatomically modern Homo sapiens, Neandertal, early Homo sapiens, and Homo erectus all lived as contemporaries over extended periods of time."³⁶  The overlap he describes is real:  9 percent of clearly ordered skull pairs run the "wrong" way, and erectus with 900 to 1,100 cc brains lived in Java long after people with 1,400 cc brains lived in Morocco.  But Lubenow's method uses the conventional dates as an order, and on those dates the ceiling rises.  He also wrote that "the significant gap between the cranial capacities of the largest australopithecine and the smallest human, fossil or living, has not been bridged."³⁶  The largest australopith brains in Du's table are 550 to 558 cc;  the smallest Dmanisi skull, which Line, Wood and Ross all count as human, is 546.³²  The gap exists only under Lubenow's own rule, that "no human fossil brain should be smaller than the smallest normal human brain today (about 775 cc for fossil adults)",³⁶ which would put four of the five Dmanisi people outside humanity;  his book predates the smallest of them, Skull 5, by nine years.  A gap made by drawing the line at a brain size cannot then be evidence about brain size.

3.  The order is real and needs a cause.  Wood accepted the trend in so many words.  After correcting for body size with the scaling line of 227 living primates, "encephalization gradually increases among humans as designated by Wood (2010, 2016b) (R2 = 0.4865, F-statistic 27.52, p = 1.57 × 10-5)",¹ and "there is definitely a trend in encephalization over radiometric time".¹  He asked "why the encephalization trend occurs in humans at all", said it "will need to be carefully re-evaluated from a creationist perspective, preferably by one trained in neuroscience", and concluded that "encephalization is no more an insurmountable problem for creationists than for evolutionists."¹  Six years earlier he had put the order of the human species into a dispersal from Babel "led by H. habilis and H. rudolfensis, specimens of which appear stratigraphically lower than any other human species", with H. sapiens "the stratigraphically-highest", and had written:  "Any model of post-Babel distribution must account for the appearance and disappearance of these human species in stratigraphic order."¹⁶  That is a dispersal model in which the smallest-brained humans lead.  It describes the order;  it does not say why the leaders had the smallest brains.

Ross, Brummel and Wood offered a reason in 2023:  "we might expect that the earliest descendants of Noah's family would exhibit a high degree of anatomical variability, some forms of which may not have persisted long enough to found a family lineage".¹⁷  They read the five Dmanisi skulls, 546 to 783 cc in one small area, as that early variety:  "Dmanisi strikingly fulfills every aspect of that prediction."¹⁷  They were candid about what their preferred account of Babel lacks:  "The disadvantage of this model is that it gives no account for either the late appearance of H. sapiens in the fossil record or the voluminous genetic and cultural evidence of H. sapiens coming to dominate and replace all other contemporaneous human forms."¹⁷

Kurt Wise, a palaeontologist trained at Harvard, had already read the rocks the same way.  From where the fossils lie, "we can deduce that H. erectus represents an early post-Babel morphology", and the hobbit, in Upper Pleistocene beds, "would date from slightly later post-Flood (and post-Babel) times."⁴⁰  His mechanism is a falling rate of change:  the fossils suggest "a high post-Flood intrabaraminic diversification rate decreasing to the present", following "pre-programmed trajectories through biological character space, the specific course of which may have been largely effected by founder effect and genetic drift in small populations following Babel."⁴⁰⁴⁰  That predicts the most variety early.  It says nothing about which way brain size should go.

Line argued in 2026, from the large Yunxian 2 skull, that "If brain size was well within the modern human range early on, then this goes against the evolutionary idea that brain size gradually increased from small H. erectus brains (e.g., the Dmanisi individuals) to modern human brains over two million years or so.  What it indicates is that there was considerable variation in brain size in 'ancient' human populations, same as today."¹²  That is the most direct young-earth reply to the trend found for this article, and it rests on a real fossil.  Part 4 weighs it.

4.  Small brains are degenerate offshoots that came first.  Rupe and Sanford turn the evolutionary sequence round:  "modern humans devolved locally into Erectus, Hobbit, and Naledi.  Those early inbred colonies were later replaced as larger (more viable) populations filled the world.  The early isolated, and inbred colonies would leave behind the earliest bones, and would soon disappear."³⁷  This is the one published young-earth mechanism that puts small brains low in the record.  Lubenow's and others' version, that the other human forms were "localized, inbred populations with congenital deformities", is rejected by Ross, Brummel and Wood:  "Inbred populations would be expected to show less genetic diversity, not more."¹⁷  Line rejects it too:  devolution "would have to have happened independently, in a relatively short time, in many different parts of the world, which seems a little too coincidental to be realistic."³⁹  Line's own suggestion is developmental:  that the robust forms "could reflect changes in development of pre-Flood and early post-Flood individuals, linked to longevity, possibly involving thyroid hormone secretion patterns".⁴¹

5.  Small brains mark apes.  Patrick Young in 2006 and Jean O'Micks in 2017 used brain size against body size to put H. habilis and H. naledi with the australopiths.⁴²³³  Young left skull 1470 out of habilis:  "So the cranial data of KNM ER 1470 will not be included as Homo habilis in this study."⁴²  Part 2 showed that O'Micks's groups are a brain-size sort.  On this view the order is partly an order of kinds:  apes below, humans above.

What has and has not been done.  The trend has been engaged directly by Wood, by Ross, Brummel and Wood, and by Line;  Lubenow argued against the order in general.  Four mechanisms that bear on the order are in print:  Wood's dispersal led by habilis, Wise's falling rate of diversification, the early variability of Ross, Brummel and Wood, and Rupe and Sanford's inbred colonies.  None of the three has been tested against the dated skulls, and none predicts the pattern the skulls show, a ceiling that rises step by step.  Early variety, from Wise or from Ross, Brummel and Wood, predicts a wide spread early, with large brains among the small.  Inbred colonies leaving the earliest bones predicts large-brained parent populations living beside them from the start.  A dispersal order says who arrived first, not why the first arrivals had the smallest brains.

Part 4: Pricing the young-earth readings

The window.  Ross, Brummel and Wood place the other human forms after the Flood and before Abraham:  "our model implies that all of these other human forms, including widespread Neandertals and H. erectus, emerged after the Flood, dispersed after Babel, and were replaced by H. sapiens prior to the time of Abraham."¹⁷  William Henry Green gave the interval from the Flood to Abraham's birth as "292 (Hebrew), 942 (Samaritan), or 1172 years (Septuagint)".⁴³  So on this model the whole sequence of human brain sizes, from Dmanisi to the Neanderthals and early sapiens, played out in 292 to 1,172 years.  How radioactive clocks could run fast enough to put three million years of dates into that span is argued in the species articles, for example Paranthropus and The South African Australopiths;  this part grants that it can and asks only about the brains.

There are two ways to read the trend inside that window.  Either the brains of one human population grew, or large and small brains were there from the start and the mix changed.

Reading A:  the brains grew.  This is Wood's "encephalization trend in humans" taken as change down the generations.  The five Dmanisi skulls average 638 cc;  the early sapiens and Neanderthals in Neubauer's sample average 1,419 and 1,438.  That is a growth of 2.2 times.  The way biologists measure the speed of such a change is in standard deviations of the trait per generation, a unit called the haldane.  Using the spread of brain size among living people, 12 percent, which is generous because Neubauer's sample mixes many populations, the change is 6.5 standard deviations.  Spread over 1,172 years at 25 years a generation, that is 0.14 standard deviations every generation;  over 292 years, 0.56.¹⁸

For scale, in D. S. Falconer's selection experiments on laboratory mice, bred each generation for body size, the response averaged 0.137 standard deviations a generation over 23 generations.⁴⁴  Philip Gingerich's survey of field studies puts natural rates at 0.01 to 0.1:  "Microevolutionary studies yield rates on the order of 10⁻¹–10⁻² haldanes calculated over intervals of historical time ranging from 10¹–10² generations."⁴⁴  So on the Septuagint, the human brain had to grow as fast as mice under deliberate breeding, for twice as many generations, in the same direction, on three continents at once.  On the Hebrew text, four times faster.  A longer window helps in proportion.  Wise bracketed the hobbit "between the Flood and ca 1000 B.C. (the time of David)",⁴⁰ and the earliest Flood date Chris Hardy and Robert Carter allow is 3386 BC.⁴⁵  If the change could run that long, about 2,400 years, the rate falls to about 0.07, inside the range of field studies, though it would then outrun Ross, Brummel and Wood's own limit of Abraham.¹⁸  That is not impossible on its face;  selection can move a trait that fast for a while.  But no young-earth writer has proposed what selected for bigger brains so hard, everywhere, and Rupe and Sanford's own account of food stress runs the other way:  "under starvation conditions those people with small brains will survive while all others will die."³⁷  The mainstream has the opposite problem.  Du's team found that on the conventional clock the change inside lineages was too slow for steady selection:  the observed increases "are always below the expected increase derived from the exponential model (except for the 1.7–1.4 Ma time period".²  On their clock the trend needs long pauses;  on a young-earth clock it needs a sprint.

Reading B:  large brains were there all along.  This is Rupe and Sanford's reading, and Line's "considerable variation in brain size in 'ancient' human populations, same as today."  It needs no growth, and Yunxian 2 is real support for it:  if its new date holds, people with 1,143 cc brains lived when the Dmanisi people had 546 to 783.  What Reading B must explain instead is a change in the mix.  In Du's table, 25 Homo skulls are wholly older than one million years.  If people with modern-sized brains had made up a tenth of the dead in those layers, the chance of finding none among 25 is 7 percent;  if a fifth, 0.4 percent.  Count Yunxian 2 as one, and the skulls allow large brains among the early dead up to about one in seven.¹⁸  Among the 36 skulls younger than 300,000 years in the table above, 30 are over 1,140 cc.  So large brains went from at most about a seventh of the dead to more than four fifths.  Reading B turns a trend in brain size into a trend in which people died where, and that still needs a cause.  Rupe and Sanford give one, the small inbred colonies leaving "the earliest bones", but on their own telling the large-brained populations were the "larger (more viable)" ones,³⁷ and larger populations would be expected to leave more bones, not fewer.

What a young-earth model would have to show.  Either a selective pressure strong enough to double brain size in a few dozen generations in every population, or a reason why, for the first part of post-Flood history, the people who died where their bones could be kept were almost all small-brained, from Georgia to Java to Kenya.  Ross, Brummel and Wood's early variability does not supply either.  They did not say which traits varied, but if brain size was among them, variety predicts large and small brains mixed early, and the early layers hold almost only small ones, Yunxian 2 the one exception.  Both readings can be tested.  Neither has been.

Part 5: The small, late brains

The facts.  The hobbit, LB1, from Liang Bua on Flores, has an endocranial volume of 426 cc;  Daisuke Kubo's team measured it by micro-CT and concluded "the original ECV of the LB1 cranium was 426 cc with only a limited degree of error, probably within ±3 cc."²²  After the 2016 re-excavation, "All skeletal remains assigned to H. floresiensis are from the pedestal deposits dated to approximately 100–60 kyr ago".²⁴  The two composite naledi skulls from the Dinaledi Chamber gave "estimates of approximately 560cc and 465cc, respectively", which "overlap entirely with the range of endocranial volumes known for australopiths."²³  Their chamber's deposits are dated "to a period between 236 ka and 335 ka."²⁵  A third skull, LES1 from the Lesedi Chamber, is about 610 ml.⁴⁶

Against the Homo trend fitted to Du's 60 Homo skulls, the hobbit sits 7.5 standard deviations below the line at its date, and the two naledi skulls 5.1 and 6.4.  For comparison, the late Javanese erectus skulls of Sambungmacan and Ngawi, also small for their date, sit 2.0 and 1.6 below.¹⁹  No Homo skull in Du's table lies more than 2.5 standard deviations from the line in either direction;  the largest departure upward is Yunxian 2 at its new date, 3.5.  Small brains also outlasted large ones elsewhere:  Paranthropus boisei, at 430 to 545 cc, is found down to 1.4 million years, long after Homo skulls of 800 to 900 cc.

What the mainstream makes of them.  Neither species is on the line that leads to living people, and the trend statements in the literature are about that line and the species around it.  Du's team stopped their table where fossil brains reach the modern range, at 0.5 million years, so neither species is in it.²  The hobbit is read as island dwarfing from a larger ancestor.  Kubo's team concluded that "evolution from early Javanese H. erectus to H. floresiensis was possible in terms of brain size", but that "some amount of the brain size reduction remains to be explained by factors other than body size scaling, posing a challenge to our current knowledge about human brain size evolution."²²  The published scaling rules reach 426 cc only from a small erectus or a habilis-sized start, never from a modern one;  that arithmetic is in Homo floresiensis: Island Dwarfing and the Small Brain.  Naledi is read as a small-brained lineage that survived in southern Africa beside larger-brained people.  So for the mainstream the two species show that brains could shrink, or stay small, on side branches.  The trend is a pattern across the lineages that left descendants, not a rule every population obeyed.  That weakens any story in which bigger brains were always favoured.  It leaves the ceiling where it was.

What young-earth writers make of them.  Most call both species human.  Wood found that every human taxon in his study "except Homo naledi has statistically significant ER", meaning naledi's brain is not significantly larger than the living-primate line predicts for its body size,¹ and he still counts it human on its skull and teeth.  Ross, Brummel and Wood suggest "the oddness of H. floresiensis and its location on Flores Island may well be explained in part by inbreeding".¹⁷  Rupe and Sanford:  "Naledi appears to be an example of this process happening within an Erectus (fully human) population.  In this light, the small cranial capacity of Naledi is not a valid reason to assign it to a new species, or deny its human status."³⁷  Line has proposed cretinism for both;  the diagnosis is tested and found wanting in Homo floresiensis: The Disease Proposals.  O'Micks alone calls naledi an ape, on the brain-size sort described in Part 2.

What they do to each side.  For the young-earth writers who count them human, these two species are the strongest evidence that brain size cannot sort humans from apes:  if Adam's descendants could have brains of 426 to 560 cc late in post-Flood history, a small brain in an early skull proves nothing about its kind.  That supports "variation, not progress", and Rupe's 2026 dropping of cranial capacity as a test follows from it.  But the same two species cut against the young-earth explanations of the order.  Rupe and Sanford's inbred colonies "would leave behind the earliest bones", yet their two clearest examples, naledi and the hobbit, are among the latest bones of all, later in the same record than the 1,375 to 1,467 cc skulls of Jebel Irhoud.  And if human brains varied this widely late, early variety cannot be what makes the early layers small-brained.  For the young-earth writers who call naledi an ape, a small-brained ape living beside large-brained humans is no puzzle;  the puzzle is that the line between them was drawn by brain size.

Neither side gets from these fossils what would most help it.  The mainstream does not get a smooth rise in every lineage.  The young-earth side does not get a large brain early.  Only Yunxian 2, if its new date stands, offers that, and Part 2 showed what one such skull does to the pattern.

Objections and responses

1. "The dates are wrong, so the order means nothing."  The order does not need the dates to be right in years, only in sequence.  The pair test asks only which skull is younger.  Inside one pile of rock, such as Olduvai or Koobi Fora, that is settled by which layer is on top.  And the young-earth writers who place these fossils read the sequence the same way:  Wood calls habilis "stratigraphically lower than any other human species", and Wise reads erectus as "an early post-Babel morphology" from where it lies.⁴⁰¹⁶  If the dates were wrong even in sequence, a rank correlation of −0.89 between brain size and date would have to be a coincidence across 94 skulls from four continents.

2. "Fossils are named by brain size, so the trend is circular."  Names were partly given by brain size, and Wood and Collard said so.⁶  That is why Part 2 dropped the names.  The order survives without them in 91 percent of clearly ordered pairs.

3. "Dates were steered to fit the expected order, as with skull 1470."  The 1470 dispute was real and both sides threw out numbers that did not fit.  Putting 1470 back at the date Lubenow defends changes the pair statistic from 91.3 to 90.7 percent.  The dispute itself was over a factor of 1.4 in age, not the hundreds that a young earth needs, and the final date rests on single crystals from several laboratories.⁴⁷

4. "Brain size has nothing to do with intelligence, so the trend means nothing."  Within living people the link between brain size and intelligence is weak, and Lubenow, Rupe and Sanford, Line and Wood all say so.  But this article tests order, not intelligence.  A pattern in which brain size rises with date needs explaining whatever brain size means.

5. "Neanderthals had bigger brains than ours, so there is no progress."  True:  Neubauer's Neanderthals average 1,450 ml and his living people 1,328.⁵  Upper Palaeolithic sapiens in Neubauer's sample average higher still, about 1,510.  Whether brains have shrunk since then is disputed.  DeSilva's team argued for a fall about 3,000 years ago;  Brian Villmoare and Mark Grabowski, re-examining the same table, concluded that "human brain size has been remarkably stable" over the last 300,000 years.⁴⁸  Either way, the top of the trend levels off after about 300,000 years.  That does not reverse the order below it.

6. "Naledi and the hobbit disprove the trend."  They break its floor, not its ceiling.  A trend in which the largest brain at each date rises can live with small brains late;  it cannot live with large brains early.

7. "Large-brained humans lived early but have not been found."  Possible, if they were rare among the early dead.  Part 4 puts a number on it:  the skulls allow at most about one in seven, against more than four in five later.  That shift in the mix still needs a cause.

8. "It is only body size."  About half of it is.  The rise from erectus to sapiens is not.

9. "Microcephaly or cretinism explains the small fossil brains."  Line himself counts "too many specimens with small cranial capacity for pathology to be an explanation for all of them" among erectus.³⁹  Disease could make a small brain anywhere in the record;  it cannot explain why the small brains are low in it.

10. From the other side:  "Young-earth writers have no model, so the trend proves evolution."  The trend is what descent with modification predicts, and it is strong evidence of a real sequence in time.  But young-earth writers do have models of dispersal and early variety, and Wood has said plainly that the trend needs a creationist explanation.  An unexplained pattern is a burden on a model, not a refutation of it.  Wood's own judgement, that encephalization is "no more an insurmountable problem for creationists than for evolutionists",¹ is fair only in one sense:  the mainstream must explain why the rise was so slow, the young-earth side why it was so fast.

Issues and unanswered questions

1. The modern mean rests on a paper not read here. Beals, Smith and Dodd 1984, Current Anthropology 25:301–330, is cited through De Caro 2024 and DeSilva 2023. The primary is on JSTOR. Whether the ±78 is a spread of individuals or of group means is not known from the secondary sources; it is almost certainly the second.
2. De Miguel and Henneberg 2001, the largest older compilation, was not obtained. Matzke reports from it that date accounts for 89 percent of the variation in cranial capacity and taxa for 5 percent, a result close to this article's, but that figure is unchecked.
3. Yunxian's date and volume. The 1.77 million years rests on one method in one 2026 paper, on samples about 75 m from where the crania lay in a layer traced between trenches. The 1,143 cc is a reconstruction of a crushed skull, and Du's team flagged the Yunxian volume as unreliable. Both could move.
4. Mixed laboratories. Du's volumes are as each original author published them; Neubauer's are virtual casts from one laboratory. Neubauer's ages are single figures. A table built in one laboratory across the whole span does not exist.
5. Body mass. Fossil body masses come from regressions on joint sizes and carry uncertainties of about a quarter. The two estimates here, from Wood's taxon means and Will's specimens, agree on "about half", but neither is precise.
6. The between-and-within split depends on how many species one counts and on the method. Du's 64 and 88 percent and this article's 53 and 69 percent are both reasonable and neither is a fact about the fossils alone.
7. No young-earth model has been computed. The questions in Part 4 can be answered: a demographic model of post-Flood dispersal that says how often small- and large-brained people should be buried at each stage, or a selection model that says what doubled brain size in a few dozen generations. None has been published.
8. LES1, the 610 ml naledi skull, comes from the Lesedi Chamber, which has no published date of its own.
9. Gould 1974, quoted by Jue for the claim that erectus and modern humans have much larger brains than their bodies predict, was not checked.

Verdict

The trend is real.  Across 94 dated skulls, and again across 283 in a second compilation, brain size rises with date by about half again every million years.  The order holds with the species names removed, with doubtful skulls removed, in one region, inside single sequences of rock, and with the most disputed date restored.  Brain size was used in naming Homo, and one young-earth paper's line between ape and human is a pure brain-size cut, but neither touches the order by date.  About half the rise follows body size;  the later half does not.

The young-earth side has conceded the pattern and not explained it.  Wood, Wise, Ross, Brummel and Wood, Rupe and Sanford, and Line all read these fossils as a sequence in post-Flood history.  Wood says the trend is real and needs a creationist cause.  The mechanisms in print, dispersal led by habilis, falling diversification, early variety, inbred colonies first, describe or allow the sequence without predicting a rising ceiling.  Priced, the lineage reading needs selection as strong as laboratory breeding, sustained in every population, and the replacement reading needs large-brained people to be at most about one in seven of the early dead and four in five of the late.

The small, late brains complicate both sides, unequally.  Naledi and the hobbit show that brains could stay small or shrink on side branches, so the mainstream trend is a pattern, not a law.  For young-earth writers who count them human they settle that brain size cannot sort humans from apes, but they are the latest of the small brains, not the earliest, which is the opposite of what the inbred-colony reading says.  The one real challenge to the ceiling is Yunxian 2, if its 2026 date holds.  It is the find that Line's "variation, same as today" predicts, and if more like it turn up in early layers, the young-earth reading will have its evidence.  So far there is one.

Sources