Are human brains shrinking? Every measured skull, 7 million years

01 Key findings

1. Skulls dated to the last 100 years average 1,297 cc; skulls older than 11,700 years average 1,458 cc (161 cc apart; 95% interval 134 to 189). Both sides of the debate accept these two averages.DeSilva et al. 2023 file; IQ Metrics calculation
2. Almost half the skulls (48.8%) are from the last 100 years, from seven collections; only 24 (2.8%) are older than 2 million years. The 2021 file had 58.7% in its last 100 years.IQ Metrics count from the three data files
3. Without the last-100-year skulls, the Holocene average is 1,430 cc against 1,458 cc before it: 28 cc lower, but not clearly different from zero (95% interval -8 to +64).IQ Metrics calculation (Welch interval)
4. The 70 skulls dated 4,200 to 100 years ago average 1,381 cc, 78 cc under the Pleistocene average (95% interval 33 to 122): a smaller but visible drop that comes before the last century.IQ Metrics calculation
5. The break date is not stable: about 3,400 years ago on the 2021 file, about 760 on the revised file, each with a wide bootstrap interval. The 2022 critique found no significant break at all.IQ Metrics calculation, simple; Villmoare and Grabowski 2022
6. Brains grew about 4.0x in 7 million years: 365 cc in Sahelanthropus, 440 cc in australopiths, 932 in Homo erectus, 1,388 in Neanderthals, 1,458 in Late Pleistocene H. sapiens.IQ Metrics calculation
7. Volume is a weak guide to intelligence: across 88 studies the correlation is r = 0.24, about 6% of the differences between people, and a bigger Neanderthal average (1,388 cc) says nothing about how they thought.Pietschnig et al. 2015
8. In living adults the trend points the other way: intracranial volume by MRI is 6.6% larger for those born in the 1970s than the 1930s (1,321 against 1,234 mL).DeCarli et al. 2024

02 Statistics to quote

03 What do the key terms mean?

Endocranial volume (ECV)

The space inside the braincase, measured on a skull or from a scan, in cubic centimetres (cc, the same as millilitres). It is a little larger than the brain it held.

Cranial capacity

The older name for endocranial volume. DeSilva et al. used published volumes in cc or mL and not brain weights in grams, because the two are not equivalent ^[3].

Hominin

A member of the human family tree after the split from the chimpanzee line: Sahelanthropus, Australopithecus, Paranthropus and Homo.

Pleistocene and Holocene

The Pleistocene is the ice-age epoch, ending 11,700 years ago; the Holocene is the warm epoch since.

Change point

A date at which a trend changes slope or level. DeSilva et al. found one in H. sapiens skull size about 3,000 years ago ^[3].

Encephalization

Brain size relative to what is expected for body size. Our figures are absolute volumes; body mass is not in these files.

Intracranial volume (MRI)

The volume inside the skull of a living person, measured on a scan. We compare it with skull volumes only as a trend, never as the same measurement.

Bootstrap interval

A range found by redrawing the skulls with replacement many times and refitting. It shows how much a result depends on which skulls are in the sample.

04 How big were the brains of our human ancestors?

The earliest hominin skull measured, Sahelanthropus tchadensis from about 7 million years ago, held about 365 cc: level with the 367 cc average of 60 chimpanzee skulls. That is the starting line of human brain evolution: 360 to 370 cc for Sahelanthropus ^[9], 300 cc for Ardipithecus ramidus at 4.4 million years, 367 cc for Australopithecus anamensis at 3.8 million ^[3]. The place of Sahelanthropus in the human family is itself disputed ^[10].

Australopiths, the next group, averaged 440 cc across 18 skulls (range 367 to 555), about 20% above a chimpanzee and roughly the size they kept for about 1.5 million years, from 3.5 to 2.0 million years ago ^[3]. Paranthropus, the heavy-jawed side branch, averaged 522 cc (19 skulls). For the living apes, see how chimpanzees compare with human children, gorilla IQ and our data on neurons in 222 species.

Cast of the cranium of Sahelanthropus tchadensis (TM 266-01-060-1), about 7 million years old: about 365 cc, the size of a chimpanzee skull.

Photo: Didier Descouens, CC BY-SA 4.0, via Wikimedia Commons

The Dmanisi cranium D4500 and its lower jaw D2600, 1.77 million years old: 546 cc, the smallest Homo erectus skull in our data.

Photo: Gerbil, CC BY-SA 4.0, via Wikimedia Commons

Figure 1

Every measured skull: how big, and how long ago (851 skulls, 7 million years)

Each dot is one skull: its age across (a log scale, each step ten times the last, older to the right) and its endocranial volume up. Volume rises from about 365 cc at 7 million years to about 1,450 cc, and the cloud of recent skulls is the last 415 dots at the far left. Earliest = Rudapithecus, Sahelanthropus and Ardipithecus; Australopiths include Paranthropus; flores. = Homo floresiensis; sapiens = Homo sapiens.

EarliestAustralopithsEarly Homo, erectusheidelbergensisnaledi, flores.Neanderthalssapiens

Endocranial volume (cc)

200600100014001800

TM 266-01-060D4500LES 1La Ferrassie

1001k10k100k1M10M

Years ago (log scale)

Skulls dated to the last 100 years are shown at 100 years ago, as in the source files. Chimpanzees have no fossil age and are not on this chart (60 skulls, mean 367 cc).

Our dataset holds one row per measured skull, built from three open files: 94 fossil crania in Du et al. 2018 ^[1], 987 rows in DeSilva et al. 2021 ^[3] and 837 in their 2023 revision ^[5], plus 60 chimpanzee skulls from a 2026 data release ^[11]. A skull that appears in two files is one row; the 2021 and 2023 files share 675 of their rows.

Figure 2

How much skulls vary inside each kind of human ancestor

Each dot is one skull. The box is the middle half of the skulls, the line the median, the whiskers the smallest and the largest. Even inside one group the spread is 400 to 800 cc, wider than the gap between neighbouring groups.

Middle halfMediansmallest to largestskull

Chimpanzee (living)60 skulls360 cc

Australopithecus18 skulls429 cc

Paranthropus19 skulls500 cc

Early Homo9 skulls673 cc

Homo erectus54 skulls912 cc

Homo heidelbergensis27 skulls1230 cc

Neanderthals26 skulls1355 cc

H. sapiens, over 11,700 years ago136 skulls1464 cc

H. sapiens, 11,700 to 100 years ago139 skulls1422 cc

H. sapiens, last 100 years415 skulls1283 cc

200600100014001800

05 When did human brain size triple?

Between about 2 million and 430,000 years ago. The average rose from 440 cc in australopiths to 678 cc in early Homo and 932 cc in Homo erectus, and the first skulls past three times the australopith average (1,318 cc) are three Atapuerca skulls (Arsuaga et al. 2014), dated about 430,000 years ago in the files (1,334 to 1,437 cc) ^[5]. Overall: about a four-fold increase over 7 million years ^[9].

The rise was not smooth. A change-point fit on 985 skulls found the rate of increase jumping at 2.10 million years ago (±0.07) and again at 1.49 million (±0.14), around the first Homo erectus ^[3]. Du et al. found that 64% or 88% of the gradual rise, depending on the taxonomy, came from change inside lineages rather than from larger-brained species replacing smaller-brained ones ^[1], and a 2024 phylogenetic study agrees that relative brain size rose within species across 7 million years ^[10]. The smallest Homo erectus, D4500 from Dmanisi, Georgia, holds 546 cc ^[14]; the Homo erectus average is 932 cc against 1,244 cc for Homo heidelbergensis.

Not every lineage grew. Homo naledi, deposited in South Africa between 236,000 and 335,000 years ago, had skulls of about 465, 560 and 610 cc, inside the australopith range, with LES1 the largest ^{[13] [12]}. Homo floresiensis from Flores was about 380 to 426 cc depending on the measurement ^{[15] [12]}.

The Homo naledi holotype, DH1 (A to K: the cranial and jaw pieces). Its volume is about 560 cc; the species runs from 465 to 610 cc.

Photo: Lee Roger Berger research team, CC BY-SA 4.0, via Wikimedia Commons

A cast of the Homo floresiensis skull. Papers give the type skull, LB1, 380 to 426 cc; our data hold 419.

Photo: Rama, CC BY-SA 3.0, via Wikimedia Commons

Figure 3

Four times a chimpanzee: average skull volume, smallest to largest

Average endocranial volume in cubic centimetres. A Late Pleistocene human skull holds 4.0 times a chimpanzee's volume and 3.3 times an australopith's; Neanderthals and Late Pleistocene H. sapiens sit within 70 cc of each other. Species with 1 to 4 skulls are single measurements, not averages of a population.

Ardipithecus ramidus (1 skull, 4.4 Ma)300 cc

Sahelanthropus tchadensis (1 skull, 7 Ma)365 cc

Chimpanzee, living (60 skulls)367 cc

Homo floresiensis (1 skull)419 cc

Australopithecus (18 skulls)440 cc

Paranthropus (19 skulls)522 cc

Homo naledi (3 skulls)545 cc

Early Homo (9 skulls)678 cc

Homo erectus (54 skulls)932 cc

Homo heidelbergensis (27 skulls)1,244 cc

H. sapiens, last 100 years (415)1,297 cc

Neanderthals (26 skulls)1,388 cc

H. sapiens 11,700 to 100 years ago (139)1,430 cc

H. sapiens over 11,700 years ago (136)1,458 cc

Figure 4

Species by species: the average and the smallest to largest skull

The dot is the average volume, the bar runs from the smallest to the largest skull measured, oldest species at the top. Averages rest on 3 to 54 skulls for fossil species and on hundreds for H. sapiens; species with a single skull are in the next figure.

A. afarensis (n = 6)424 (372–522)

A. africanus (n = 9)459 (400–555)

P. boisei (n = 10)486 (400–545)

P. robustus (n = 8)581 (450–800)

H. habilis (n = 6)629 (509–777)

H. rudolfensis (n = 3)776 (750–825)

H. naledi (n = 3)545 (465–610)

H. erectus (n = 54)932 (546–1800)

H. heidelbergensis (n = 26)1238 (880–1450)

Neanderthals (n = 26)1388 (1065–1740)

H. sapiens, over 11,700 y ago (n = 136)1458 (1100–1775)

H. sapiens, 11,700 to 100 y ago (n = 139)1430 (1030–1861)

H. sapiens, last 100 y (n = 415)1297 (975–1776)

300700110015001900

06 Did Neanderthals have bigger brains than we do?

Yes, on average: 26 Neanderthal skulls average 1,388 cc against 1,304 cc for 455 modern H. sapiens skulls under 1,000 years old in the same file, 85 cc more (our 95% interval 15 to 155). Neanderthal brain size matches that of Late Pleistocene H. sapiens, not the modern average.

La Ferrassie 1, a Neanderthal skull: 1,640 cc in our data, about 70,000 years old.

Photo: Wikimedia Commons user 120, CC BY 2.5, via Wikimedia Commons

The skull of the Old Man of Cro-Magnon (Cro-Magnon 1), about 27,700 years old: 1,730 cc, among the largest in our data.

Photo: Wellcome Collection, CC BY 4.0, cropped, via Wikimedia Commons

Neanderthals who lived in the last glacial (under 115,000 years ago, 14 skulls) average 1,459 cc, effectively the same as the 1,458 cc of Late Pleistocene H. sapiens ^[5]. A 2023 review lists a Neanderthal average of 1,415 cc (25 skulls) and 1,499 cc for Pleistocene H. sapiens (29 skulls), from older measurements ^[9]: different skulls, the same picture. The largest Neanderthal in the files, Amud, is about 1,740 cc.

How smart were Neanderthals? Skull volume cannot say. It does not include body size, wiring or learning, and in living people it explains about 6% of the differences in IQ scores ^[8]; see brain size and IQ: what a correlation of 0.24 means. A bigger average is a fact about skulls.

07 What is the average human brain size, in cubic centimetres?

Roughly 1,300 to 1,370 cc, depending on the sample: 1,304 cc for 455 museum and anatomical skulls in the revised file, 1,349 cc for 5,288 skulls in a global survey, 1,371 cc in MRI scans of 2,717 adults in the United States and Europe ^[5]. A cubic centimetre is a millilitre. A cranial capacity is not the same as a brain weight, which is why published averages differ.

The choice matters for the shrinking-brain question. A 2022 critique called the 1,297 cc average of the 2021 file too low against 1,340 to 1,460 cc in other published estimates ^[4]; the authors answered that the same sources give 1,308 to 1,392 cc ^[5]. A 2024 commentary pooled three samples and a revised file into 10,363 skulls and brains at 1,341 cc ^[6].

Figure 5

The size of the drop depends on which "modern" average you pick: 87 to 154 cc

The Late Pleistocene H. sapiens average in the revised file (1,458 cc, n = 136) against six published modern averages. Same earlier skulls, different modern reference: the gap runs from 87 to 154 cc.

Modern averageLate Pleistocene average

Our 455 museum and anatomical skulls1,3041,458

Tattersall 2023, "contemporary" (more than 500)1,3301,458

Medical, United States (3,399 brains, converted)1,3341,458

De Caro 2024, all sources combined (10,363)1,3411,458

Beals et al. 1984, global survey (5,288 skulls)1,3491,458

MRI, United States and Europe (2,717)1,3711,458

1,2501,3001,3501,4001,4501,500

Average volume (cc)

08 Are human brains shrinking?

Skulls from the last 100 years average 1,297 cc, 161 cc under skulls older than 11,700 years (1,458 cc), but whether that is shrinking or sampling is disputed. Nineteen earlier studies report a decrease averaging 8.5% ^[5]; a 2022 reanalysis found none once the sample was rebalanced ^[4].

The claim (2021)

DeSilva, Traniello, Claxton and Fannin put 985 dated skulls through a change-point analysis and found that human brain reduction was “surprisingly recent, occurring in the last 3,000 years” ^[3]. The fitted break was at 0.003 ± 0.001 million years ago, and the rate of change after it was 50 times the rate of the earlier rise. They called the result conservative and proposed that the externalization of knowledge in groups, as in ant colonies, may explain it ^[3].

The critique (2022)

Villmoare and Grabowski noted that 578 of the 987 specimens, more than half, were dated to the final 100 years, from two sources, and that only 23 skulls covered the 5,000 to 1,000 years the claim depends on. Their modern average, 1,297 cc, sat below published ones. They reduced the data to 100-year means, tested the last 300,000 and the last 30,000 years, and found no change point (Davies test p = 0.739 and 0.259) and a flat trend (p = 0.754): “human brain size has been remarkably stable over the last 300 ka” ^[4].

The reply (2023) and the re-check (2024)

DeSilva and seven colleagues reaffirmed the reduction. They removed the Morton collection, which they called problematic, added 94 skulls from the Terry collection, 50 from India, 13 dated 1,000 to 5,000 years ago and three measured in South Africa, dropped four juvenile Neanderthals and a duplicate, and argued that binning dilutes the variation. With simple t-tests they found a drop of 159 cc (or 117 cc against a larger modern reference) after 3,000 years (t = 12.81) and 10.8% overall ^[5]. In 2024 De Caro pooled the revised file with three large samples, agreed the reduction in the last 5,000 years looked real, but argued one 25-skull sample from Afalou, Algeria, should come out and that the fit explains little of the variation ^[6].

Figure 6

How the 3,000-year claim travelled: claim, critique, reply, re-check

Each step in time order, from the earlier work the claim rests on to the latest check. Steps are evenly spaced; the year is printed above each dot.

New data or a new analysisA critique or a checkEarlier estimate

The 3,000-year claim, 2021 to 2024985 dated skulls in the first file

1. 2021DeSilva et al.: a change point 3,000 years agoPiecewise fit to 985 dated skulls; reduction "surprisingly recent"
2. 2022Villmoare and Grabowski: no change point578 of 987 skulls in the final 100 years; no break once rebalanced
3. 2023DeSilva et al. reply with a revised fileMorton skulls out, about 160 skulls added; a drop of 159 or 117 cc
4. 2024De Caro re-checks the replyConfirms a drop; questions one 25-skull sample

Earlier work the claim builds onas compiled by DeSilva et al.

1. 1934von Bonin: 3.9% smaller in EuropeOver 10,000 or 20,000 years
2. 1984Beals et al.: 9.0% on 5,288 skullsOver the last 100,000 years
3. 1988Henneberg: 13.5%About 12,800 skulls; since 20,000 years
4. 1997Ruff et al.: 11.1%Over 35,000 years
5. 2004Henneberg: 7.3% over the last 3,000 yearsMore than 14,000 skulls

Both sides use the same skulls and mostly agree on the averages. They disagree on whether skulls from a handful of collections can stand for the world’s people over time. The size of the claimed drop also depends on the choice of “modern” average, as Figure 5 shows.

Figure 7

Nineteen earlier studies of a shrinking brain: from 3.9% to 13.5%

Percentage decrease reported by each study, as compiled in Table S1 of the 2023 reply (Wiercinski gave 5.6 to 5.8, shown as 5.7; Brown gave two figures). The authors report an average of 8.5%; recomputed from the table it is 8.45% across the 19 with a figure. The two DeSilva et al. papers are highlighted.

Henneberg, 198813.5%

DeSilva et al., 202111.5%

Ruff et al., 199711.1%

DeSilva et al., 202310.8%

Stibel, 202210.7%

Henneberg & Steyn, 199510.0%

Henneberg, 199810.0%

Brown, 1992 (Australia)9.5%

Beals et al., 19849.0%

Schwidetzky, 19768.9%

Balzeau et al., 20138.7%

Henneberg & Steyn, 19938.4%

Henneberg, 20047.3%

Wu et al., 20077.2%

Brown & Maeda, 20047.0%

Wiercinski, 19795.7%

Stibel, 20215.4%

Brown, 1992 (East Asia)5.0%

Liu et al., 20144.2%

von Bonin, 19343.9%

09 Where do the skulls in the shrinking-brain data come from?

Mostly from the last 100 years: 415 of the 851 skulls in our analysis set (48.8%), and 578 of 985 in the 2021 file (58.7%). They come from seven collections in the United States, Australia, China, India and England, dated in the files to “100 years ago”. Only 24 skulls (2.8%) are older than 2 million years, and the 300,000 years before the Holocene hold 188 (22.1%) ^{[3] [5]}.

Figure 8

Every skull in our analysis set, one square each: 851 skulls, mostly Homo sapiens

Oldest first, from the top left. 690 of the 851 squares are Homo sapiens, and 415 of those are dated to the last 100 years. Hover a square for the skull.

Miocene ape (Rudapithecus) 2Sahelanthropus and Ardipithecus 2Australopithecus 18Paranthropus 19Early Homo (habilis, rudolfensis) 9Homo erectus (incl. Dmanisi) 54Homo heidelbergensis (incl. Harbin) 27Homo naledi and H. floresiensis 4Neanderthals 26Homo sapiens 690

851 skullsone square = one skull

Figure 9

Where the skulls come from: half of them are from the last 100 years

Skulls per period, coloured by group. 415 of 851 skulls (48.8%) are dated to the last 100 years, 188 (22.1%) to the 300,000 years before the Holocene and only 24 (2.8%) are older than 2 million years.

Miocene ape (Rudapithecus)Sahelanthropus and ArdipithecusAustralopithecusParanthropusEarly Homo (habilis, rudolfensis)Homo erectus (incl. Dmanisi)Homo heidelbergensis (incl. Harbin)Homo naledi and H. floresiensisNeanderthalsHomo sapiens

242M+

852M-300k

188300k-12k

13912k-100

415100

The window the claim turns on is thin. Of 673 H. sapiens skulls up to 50,000 years old in the revised file, 218 are older than 1,000 years, and only 30 are dated 1,000 to 5,000 years ago; the 2021 file had 48, of which 36 came from two sites dated to about 1,000 years. The critics counted 23 in their window ^[4]; the count depends on where the 1,000-year sites are placed.

Figure 10

From 1,918 rows to the 30 skulls that cover the years the claim turns on

Rows in the three data files; different skulls once duplicates are matched; the analysis set (the authors' 2023 version plus the 14 skulls only Du et al. list); H. sapiens in the last 50,000 years; of those older than 1,000 years; and those dated 1,000 to 5,000 years ago, the window that carries the 3,000-year claim.

Rows across the three data files1,918all

60.6% of the stage above

Different skulls after matching1,16360.6% of all

73.2% of the stage above

Analysis set85144.4% of all

79.1% of the stage above

H. sapiens, last 50,000 years67335.1% of all

32.4% of the stage above

Older than 1,000 years21811.4% of all

13.8% of the stage above

1,000 to 5,000 years old301.6% of all

Figure 11

The 987 volumes of the original data: one tall peak around 1,300 cc

All 987 rows of the 2021 file in 100 cc bands (it counts 985 dated skulls in its model). The tall peak at 1,200 to 1,400 cc is mostly recent skulls; the 49 rows under 700 cc are almost all australopiths and early Homo.

1718167613135014119121015110245133Median: 1,310 cc

200 cc400 cc600 cc800 cc1000 cc1200 cc1400 cc1600 cc1800 cc

Endocranial volume (cc)

10 What happens when we redo the analysis?

The biggest step down is the last one. Using the authors’ own periods (we reproduce their Table S2 exactly), H. sapiens skulls average 1,458 to 1,475 cc from 50,000 to 11,700 years ago, 1,396 cc at 4,200 to 1,000 years (n = 30), 1,370 cc at 1,000 to 100 years (n = 40) and 1,297 cc in the last 100 years (n = 415). This is our calculation from their file.

Figure 12

Average skull volume of H. sapiens, period by period: lower in each later period

Revised 2023 file, H. sapiens up to 50,000 years ago: the average (dot) and its 95% interval (bar) for each period, with the number of skulls. The last 100 years average 1,297 cc (n = 415), 161 cc below the 136 skulls older than 11,700 years (95% interval 134 to 189).

50,000 to 29,000 years ago (n = 15)1475 (1403–1546)

29,000 to 11,700 years ago (n = 104)1458 (1430–1486)

11,700 to 8,300 years ago (n = 41)1509 (1459–1559)

8,300 to 4,200 years ago (n = 28)1439 (1379–1498)

4,200 to 1,000 years ago (n = 30)1396 (1340–1452)

1,000 to 100 years ago (n = 40)1370 (1315–1424)

Last 100 years (n = 415)1297 (1283–1312)

1250135014501550

Leave out the last 100 years and the Holocene is 1,430 cc against 1,458 cc for the Pleistocene: 28 cc lower, with a 95% interval of -8 to +64, so not clearly different. But the 70 skulls dated 4,200 to 100 years ago average 1,381 cc, 78 cc under the Pleistocene (interval 33 to 122), and the 30 skulls at 4,200 to 1,000 years are 62 lower (interval 4 to 121). The earliest Holocene period (11,700 to 8,300 years) is high at 1,509 cc, and 25 of its 41 skulls are one Algerian site, the sample De Caro asked to remove ^[6].

Figure 13

What the authors' own revision changed, period by period

Average H. sapiens volume in each period in the 2021 file (as published) and in the 2023 file (after the critique). The big moves are in the thin periods: 4,200 to 1,000 years ago rises from 1,351 to 1,396 cc as skulls are added; 1,000 to 100 years ago falls from 1,530 (n = 11) to 1,370 (n = 40).

2021 file2023 revised file

50,000 to 29,000 years ago (n = 16 / 15)1,4751,481

29,000 to 11,700 years ago (n = 104 / 104)1,458

11,700 to 8,300 years ago (n = 41 / 41)1,509

8,300 to 4,200 years ago (n = 21 / 28)1,4391,484

4,200 to 1,000 years ago (n = 48 / 30)1,3511,396

1,000 to 100 years ago (n = 11 / 40)1,3701,530

Last 100 years (n = 578 / 415)1,297

1,2501,4001,550

Average volume (cc); n = skulls in the 2021 / 2023 file

Then the break date. We fitted a continuous line with one bend to H. sapiens skulls up to 50,000 years old (IQ Metrics calculation, simple: it knows nothing about collections, body size or region). On the 2021 file the best bend is 3,400 years ago (95% bootstrap interval 300 to 5,900), close to the paper’s 3,000. On the revised file it moves to 760 years ago (730 to 8,800). Leave the last-100-year skulls out and it is still 760 years (510 to 11,500). The fitted fall from bend to the youngest skulls is 130 to 240 cc across versions. What it does not show: a date.

Figure 14

When did the break happen? It depends on which skulls you keep (IQ Metrics calculation, simple)

A two-segment line fitted to H. sapiens skulls up to 50,000 years old; the dot is the best break date and the bar its 95% interval over 1,500 resamples of the skulls. On the 2021 file the break is 3,400 years ago; on the revised file it moves to 760 years ago; the bars overlap but are wide. Intervals ignore that skulls come in collections, so they are too narrow if anything.

2021 file as published (n = 819)3370 (300–5920)

2021 file, at most 30 skulls per period (n = 819)3370 (300–7460)

2023 revised file (n = 673)760 (730–8810)

2023 file, at most 30 skulls per period (n = 673)760 (760–11480)

2023 file without the last-100-year skulls (n = 258)760 (510–11480)

04000800012000

Our reading, from the data alone: skulls under about 4,000 years old run smaller than Pleistocene skulls in these collections, by roughly 60 to 160 cc depending on the period and the reference; the data do not fix when it began, and they cannot say whether seven collections speak for everyone. Both papers’ conclusions are consistent with parts of that.

11 Why would human brains shrink?

No cause is settled. The 2021 paper says its dating does not support three older ideas: a by-product of smaller bodies, a shift to an agricultural diet, and self-domestication ^[3]. It notes that body size may have fallen several millennia before brain size, and cites Hawks (2011): a 5 kg fall in body mass would explain only 22 mL of the drop, which he found more than five times larger ^[3]. Its own proposal is group-level cognition: knowledge stored outside any one head. The 2023 reply adds that domestic mammals also lost brain volume over the last 10,000 years ^[5].

These are hypotheses about a drop whose size and timing are themselves disputed, so we do not rank them. If the drop is mostly the last 100 years, as our data suggest, changes in body size, diet and health over the 20th century are the obvious ones to test, and nothing in these files can.

12 Are human brains getting bigger again?

In one large study, yes: adults born in the 1970s have an intracranial volume 6.6% larger than those born in the 1930s (1,321 against 1,234 mL, 3,226 people scanned by MRI). The Framingham Heart Study found the rise in volume, white matter, hippocampus and cortical surface area, and suggested early-life conditions as the cause ^[7].

This does not contradict the skull data: it measures living adults in one US town, by a different method, over 40 birth years. It does show that average intracranial volume can differ between generations born decades apart, which the authors put down to early-life conditions. They note that IQ scores also rose across the 20th century, as in a century of rising IQ scores that has stalled in wealthy countries, though this study did not test the link ^[7]. For what changes in a brain across life, see at what age intelligence peaks and our data on peak ages.

Figure 15

Living adults: intracranial volume by birth decade, Framingham, 3,226 people

Average volume inside the skull of adults scanned by MRI between 1999 and 2019, adjusted for age at scan and sex. The 1930s and 1970s values and their 95% intervals are in the paper; the three decades between are our sums of its published increments. The 1970s are 87 mL (6.6%) above the 1930s.

Born in the 1930s (n = 532)1234 mL

Born in the 1940s (n = 777)1264 mL

Born in the 1950s (n = 904)1285 mL

Born in the 1960s (n = 872)1304 mL

Born in the 1970s (n = 136)1321 mL

1200125013001350

13 Does a bigger brain mean more intelligence?

Only weakly, among living people: across 88 imaging studies of 148 samples and more than 8,000 people the correlation between brain volume and IQ is r = 0.24 (R-squared 0.06), and the authors warn against treating brain size as a proxy for intelligence differences ^[8]. Squared, that is about 6% of the variation; 94% lies elsewhere. Our news page on the correlation works through the disputes about that number.

For this study it means two things. A 100 or 160 cc difference in average skull volume, about 8% to 11%, says nothing by itself about any ability, and nothing here compares people. And within our own data a skull is a skull: no skull has a test score, and even a living person’s score is a range, as the margin of error shows. Related reading: what MRI scans of gifted brains show, famous brains weighed, how smart were dinosaurs and what happens to the brain in space.

14 How accurate are the skull measurements?

Usually within a few tens of cc, but not always. For the 80 skulls listed in both Du et al. 2018 and the DeSilva files, the median gap is 12.9 cc and 70.0% are within 25 cc; 11.2% differ by more than 50 cc and 3 by 100 cc or more ^{[1] [3]}. Where Du et al. had several observers measure one skull, the range across observers was a median of 45.1 cc (up to 208.5 cc, for OH 7).

Figure 16

The same skull, two answers: where the compilations disagree most

Of the 80 skulls listed in both Du et al. 2018 (average of 1 to 6 observers) and the DeSilva et al. files, the median gap is 12.9 cc and 70.0% are within 25 cc; 3 differ by 100 cc or more. The 12 biggest gaps are shown.

Du et al. 2018DeSilva et al. 2021 / 2023

ZKD V (H. erectus)1,1401,300

Yunxian (H. erectus)1,0501,150

Sangiran 12 (H. erectus)9631,059

Buia (H. erectus)900995

Ceprano (H. heidelbergensis)1,0571,150

OH 7 (H. habilis)685777

Sts 19/58 (A. africanus)436509

KNM-ER 3733 (H. erectus)782848

Trinil 2 (H. erectus)901940

KNM-ER 1590 (H. rudolfensis)788825

Type 2 (A. africanus)457494

SK 1585 (P. robustus)494530

4008001,200

Endocranial volume (cc)

For the shrinking-brain question this error is small: the claimed drop is 100 to 160 cc on averages of hundreds of skulls. It matters more for single species: one skull, such as Sahelanthropus, carries a range of 360 to 370 cc, Homo floresiensis 380 to 426 cc between papers, and the two Dmanisi skulls D2280 and D2282 are listed at 730 and 650 cc in one paper ^[12], 775 and 655 in another ^[3] and 650 and 780 in a 2026 dataset ^[11]. We keep every value in the CSV and say which one the page uses.

15 What is new in 2024 to 2026?

As of 3 November 2026 we found no paper that re-tests the 3,000-year claim after the 2024 commentary ^[6]. Related work: a 2024 PNAS analysis found that relative brain size rose within species across about 7 million years and accelerated in more recent lineages ^[10]; a 2023 review of endocranial volumes calls the recent shrinkage within Homo sapiens “dramatic” ^[9]; and the Framingham cohort above appeared in 2024 ^[7].

In 2025 two papers placed the Harbin cranium, whose endocranial capacity was estimated at about 1,420 ml in 2021 ^[16], in a Denisovan population, by proteins ^[17] and by mitochondrial DNA from dental calculus ^[18]. In June 2026 a dataset of 60 chimpanzee and 48 hominin volumes for a new Sts 25 estimate was released openly ^[11]; we use its chimpanzee values and have not read the article it accompanies.

16 Every specimen in our dataset

Table 1

Every skull in the analysis set: 330 named skulls and 13 collections of recent skulls

Oldest first. Years ago = the date in the source file (three significant figures). Volume = endocranial volume in cc as used in our analysis (the 2023 file where the skull is in it, otherwise Du et al.'s average). A collection row gives the average of its skulls; the CSV lists every skull with all values and references.

Skull or collection	Species	Years ago	Volume (cc)	Skulls	Reference in the source file
RUD 200	Rudapithecus hungaricus	9,850,000	234	1	Gunz et al., 2020
RUD 77	Rudapithecus hungaricus	9,850,000	326	1	Begun
TM 266-01-060	Sahelanthropus tchadensis	7,000,000	365	1	Falk
ARA-VP-6/500	Ardipithecus ramidus	4,400,000	300	1	Suwa et al., 2009
MRD-VP-1/1	Australopithecus anamensis	3,800,000	367	1	Haile-Selassie et al., 2019
AL 333-105	Australopithecus afarensis	3,220,000	372	1	Holloway et al. 2004
AL 162-28	Australopithecus afarensis	3,200,000	393	1	Holloway, 1981
AL 333-45	Australopithecus afarensis	3,200,000	488	1	Gunz et al., 2020
AL 288-1	Australopithecus afarensis	3,180,000	388	1	Gunz et al., 2020
AL 822-1	Australopithecus afarensis	3,100,000	382	1	Gunz et al., 2020
AL 444-2	Australopithecus afarensis	3,000,000	522	1	Gunz et al., 2020
Taung-1	Australopithecus africanus	2,800,000	428	1	Holloway et al. 2004
MLD 1	Australopithecus africanus	2,650,000	510	1	Holloway et al., 2004
MLD 37/38	Australopithecus africanus	2,650,000	430	1	Holloway et al., 2004
Bouri	Australopithecus garhi	2,500,000	450	1	Holloway et al., 2004
KNM-WT 17000	Paranthropus aethiopicus	2,500,000	410	1	Holloway et al., 2004
Omo L388y-6	Paranthropus boisei	2,380,000	427	1	Holloway et al., 2004
Stw 505	Australopithecus africanus	2,350,000	555	1	Holloway et al., 2004
Omo 323-1976-896	Paranthropus boisei	2,230,000	490	1	Brown et al., 1993
Sts 19/58	Australopithecus africanus	2,100,000	436	1	Holloway et al., 2004
Sts 5	Australopithecus africanus	2,100,000	485	1	Holloway et al., 2004
Sts 60	Australopithecus africanus	2,100,000	400	1	Holloway et al., 2004
Sts 71	Australopithecus africanus	2,100,000	428	1	Holloway et al., 2004
Type 2	Australopithecus africanus	2,100,000	457	1	Holloway et al., 2004
DNH 155	Paranthropus robustus	2,000,000	450	1	Martin et al., 2020
SK 46	Paranthropus robustus	1,980,000	800	1	Broom et al. 1952
SK 48	Paranthropus robustus	1,980,000	750	1	Broom et al. 1952
SK 54	Paranthropus robustus	1,980,000	500	1	Holloway et al. 2004
SK 859	Paranthropus robustus	1,980,000	465	1	Holloway et al. 2004
MH 1	Australopithecus sediba	1,980,000	420	1	Berger et al., 2010
TM 1517	Paranthropus robustus	1,900,000	625	1	Schepers 1946
KNM-ER 1805	Homo habilis	1,890,000	582	1	Holloway et al., 2004
KNM-ER 1813	Homo habilis	1,890,000	509	1	Holloway et al., 2004
OH 24	Homo habilis	1,880,000	597	1	Tobias, 1991
KNM-ER 1470	Homo rudolfensis	1,880,000	752	1	Holloway et al., 2004
KNM-ER 3732	Homo rudolfensis	1,880,000	750	1	Holloway et al., 2004
KNM-ER 13750	Paranthropus boisei	1,870,000	475	1	Holloway et al., 2004
KNM-ER 23000	Paranthropus boisei	1,870,000	491	1	Brown et al., 1993
KNM-ER 1590	Homo rudolfensis	1,850,000	825	1	Holloway et al., 2004
KNM-ER 407	Paranthropus boisei	1,850,000	510	1	Holloway et al., 2004
OH 5	Paranthropus boisei	1,820,000	523	1	Holloway et al., 2004
OH 7	Homo habilis	1,800,000	777	1	Tobias, 1991
D2280	Homo erectus	1,770,000	775	1	Rightmire et al., 2006
D2282	Homo erectus	1,770,000	655	1	Rightmire et al., 2006
D2700	Homo erectus	1,770,000	600	1	Rightmire et al., 2006
D3444	Homo erectus	1,770,000	650	1	Lordkipanidze et al., 2006
D4500	Homo erectus	1,770,000	546	1	Lordkipanidze et al., 2013
KNM-WT 17400	Paranthropus boisei	1,770,000	400	1	Holloway et al., 2004
Drimolen	Paranthropus robustus	1,750,000	530	1	Schwartz & Tattersall
Sangiran 31	Homo erectus	1,700,000	700	1	Tyler 1996
OH 16	Homo habilis	1,670,000	638	1	Tobias, 1991
SK 1585	Paranthropus robustus	1,650,000	530	1	Holloway et al., 2004
Gongwang 1	Homo erectus	1,630,000	780	1	Holloway et al., 2004
KNM-ER 3733	Homo erectus	1,630,000	848	1	Holloway, 1981
Mojokerto 1	Homo erectus	1,620,000	800	1	Anton 1997
Sangiran 4	Homo erectus	1,620,000	908	1	Holloway et al., 2004
OH 13	Homo habilis	1,600,000	673	1	Tobias, 1991
KNM-ER 406	Paranthropus boisei	1,580,000	500	1	Holloway et al., 2004
KNM-ER 732	Paranthropus boisei	1,580,000	500	1	Holloway et al., 2004
KNM-ER 3883	Homo erectus	1,570,000	804	1	Holloway et al., 2004
DAN5	Homo erectus	1,550,000	598	1	Semaw et al., 2020
KNM-ER 42700	Homo erectus	1,550,000	691	1	Spoor et al., 2007
Sangiran 38	Homo erectus	1,520,000	850	1	Indriati and Anton, 2010
KNM-WT 15000	Homo erectus	1,490,000	880	1	Begun & Walker, 1993
Konso	Paranthropus boisei	1,420,000	545	1	Holloway et al., 2004
Sangiran 10	Homo erectus	1,290,000	855	1	Holloway et al., 2004
Sangiran 12	Homo erectus	1,290,000	1,059	1	Holloway et al., 2004
Sangiran 17	Homo erectus	1,290,000	1,004	1	Holloway, 1981
Sangiran 2	Homo erectus	1,290,000	813	1	Holloway, 1981
BSN12/P1	Homo erectus	1,260,000	850	1	Semaw et al., 2020
Bukaran (SBK-1996.02)	Homo erectus	1,250,000	916	1	Grimaud-Herve et al., 2012
Grogol-Wetan (Gwn-1993.09)	Homo erectus	1,250,000	850	1	Grimaud-Herve, pers comm
OH 9	Homo erectus	1,250,000	1,067	1	Holloway et al., 2004
Sangiran IX	Homo erectus	1,250,000	870	1	Kaifu et al., 2011
Lantian	Homo erectus	1,140,000	780	1	Olivier and Tissier 1975
Buia	Homo erectus	1,000,000	900	1	Macciarelli et al., 2004
Daka	Homo erectus	1,000,000	986	1	Gilbert and Asfaw, 2008
Yunxian	Homo erectus	936,000	1,050	1	Vialet et al., 2010
Sangiran 3	Homo erectus	900,000	950	1	Holloway et al., 2004
Trinil 2	Homo erectus	900,000	940	1	Holloway, 1981
KNM-OL 45500	Homo erectus	882,000	715	1	Potts et al. 2004
Ceprano	Homo heidelbergensis	850,000	1,150	1	Bruner and Manzi, 2005
ZKD III	Homo erectus	730,000	915	1	Weidenreich, 1943
Zhoukodian Z-11	Homo erectus	730,000	1,015	1	Weidenreich, 1943
Zhoukodian Z-2	Homo erectus	730,000	1,030	1	Weidenreich, 1943
Zhoukodian Z10	Homo erectus	730,000	1,225	1	Weidenreich, 1943
Zhoukodian Z12	Homo erectus	730,000	1,030	1	Weidenreich, 1943
OH 12	Homo erectus	700,000	727	1	Holloway et al., 2004
Zhoukoudian VI	Homo erectus	602,000	850	1	Weidenreich 1943
Nanjing	Homo erectus	600,000	876	1	Liu et al., 2005
Bodo	Homo heidelbergensis	600,000	1,250	1	Conroy et al., 2000
Ngawi	Homo erectus	500,000	870	1	Holloway et al., 2004
Saldanha	Homo heidelbergensis	500,000	1,225	1	Tobias, 1971
ZKD V	Homo erectus	450,000	1,140	1	Wu et al., 2010
Arago	Homo heidelbergensis	450,000	1,166	1	Holloway et al., 2004
Atapuerca 10	Homo heidelbergensis	430,000	1,218	1	Arsuaga et al., 2014
Atapuerca 12	Homo heidelbergensis	430,000	1,228	1	Arsuaga et al., 2014
Atapuerca 13	Homo heidelbergensis	430,000	1,437	1	Arsuaga et al., 2014
Atapuerca 15	Homo heidelbergensis	430,000	1,284	1	Arsuaga et al., 2014
Atapuerca 16	Homo heidelbergensis	430,000	1,236	1	Arsuaga et al., 2014
Atapuerca 17	Homo heidelbergensis	430,000	1,219	1	Arsuaga et al., 2014
Atapuerca 2	Homo heidelbergensis	430,000	1,334	1	Arsuaga et al., 2014
Atapuerca 4	Homo heidelbergensis	430,000	1,360	1	Arsuaga et al., 2014
Atapuerca 5	Homo heidelbergensis	430,000	1,092	1	Arsuaga et al., 2014
Atapuerca 7	Homo heidelbergensis	430,000	1,144	1	Arsuaga et al., 2014
Hexian	Homo erectus	412,000	1,025	1	Wu & Dong, 1982
Ndutu	Homo heidelbergensis	400,000	1,100	1	Holloway in Rightmire, 1983
Ngandong VI	Homo erectus	344,000	1,097	1	Holloway 1983
Ngandong XII	Homo erectus	344,000	1,145	1	Grimaud-Herve et al. 2012
Swanscombe	Homo heidelbergensis	300,000	1,325	1	Holloway et al., 2004
Jebel Irhoud 1	Homo sapiens	300,000	1,305	1	Holloway et al., 2004
Jebel Irhoud 2	Homo sapiens	300,000	1,400	1	Holloway et al., 2004
Kabwe	Homo heidelbergensis	299,000	1,325	1	Holloway et al., 2004
DH1/DH2	Homo naledi	286,000	560	1	Garvin et al., 2017
DH3/DH4	Homo naledi	286,000	465	1	Garvin et al., 2017
LES 1	Homo naledi	286,000	610	1	Hawks et al., 2017
Jinniushan	Homo heidelbergensis	280,000	1,390	1	Holloway et al., 2004
Guomde	Homo heidelbergensis	272,000	1,400	1	Holloway et al., 2004
Apidima 2	Homo neanderthalensis	252,000	1,454	1	Harvarti et al., 2011
Saccopastore 1	Homo neanderthalensis	250,000	1,174	1	Bruner and Manzi, 2008
Saccopastore 2	Homo neanderthalensis	250,000	1,300	1	Holloway et al., 2004
Sale	Homo heidelbergensis	240,000	880	1	Holloway et al., 2004
Narmada	Homo heidelbergensis	236,000	1,260	1	Holloway et al., 2004
Omo 2 (Kibbish)	Homo sapiens	233,000	1,435	1	Holloway et al., 2004
Ehringsdorf	Homo heidelbergensis	230,000	1,450	1	Holloway et al., 2004
Eliye Springs (KNM-ER 11693)	Homo heidelbergensis	225,000	1,212	1	Brauer et al., 1997
Steinheim	Homo heidelbergensis	225,000	1,140	1	Prossinger et al., 2003
Petralona	Homo heidelbergensis	210,000	1,230	1	Holloway et al., 2004
Dali	Homo heidelbergensis	209,000	1,120	1	Holloway et al., 2004
Reilingen	Homo neanderthalensis	200,000	1,430	1	Holloway et al., 2004
Biache	Homo neanderthalensis	175,000	1,200	1	Holloway et al., 2004
Fontechevade II	Homo neanderthalensis	160,000	1,350	1	Holloway et al., 2004
Herto	Homo sapiens	160,000	1,450	1	Holloway et al., 2004
La Chaise	Homo neanderthalensis	150,000	1,065	1	Holloway et al., 2004
Harbin	Homo longi	148,000	1,420	1	Ji et al., 2021
Singa 1	Homo sapiens	133,000	1,550	1	Holloway et al., 2004
Krapina 2	Homo neanderthalensis	130,000	1,450	1	Holloway et al., 2004
Krapina 3	Homo neanderthalensis	130,000	1,255	1	Holloway et al., 2004
Krapina 5	Homo neanderthalensis	130,000	1,535	1	Caspari & Radovcic, 2006
Krapina 6 (Cranium E)	Homo neanderthalensis	130,000	1,205	1	Holloway et al., 2004
Lazaret	Homo neanderthalensis	130,000	1,250	1	Holloway et al., 2004
Skhul 1	Homo sapiens	130,000	1,450	1	Holloway et al., 2004
Skhul 4	Homo sapiens	130,000	1,554	1	Holloway et al., 2004
Skhul 5	Homo sapiens	130,000	1,520	1	Holloway et al., 2004
Skhul 9	Homo sapiens	130,000	1,590	1	Holloway et al., 2004
Yingkou	Homo sapiens	130,000	1,390	1	Holloway et al., 2004
Laetoli 18	Homo sapiens	120,000	1,367	1	Holloway et al., 2004
Xuchang 1	Homo erectus	115,000	1,800	1	Li et al., 2017
Ngandong (Solo 9)	Homo erectus	110,000	1,135	1	Holloway et al., 2004
Ngandong 1	Homo erectus	110,000	1,172	1	Holloway et al., 2004
Ngandong 13	Homo erectus	110,000	1,231	1	Holloway et al., 2004
Ngandong 14	Homo erectus	110,000	1,090	1	Holloway et al., 2004
Ngandong 6	Homo erectus	110,000	1,251	1	Holloway et al., 2004
Ngandong 7	Homo erectus	110,000	1,028	1	Balzeau et al., 2003
Tabun 1	Homo neanderthalensis	110,000	1,271	1	Holloway et al., 2004
Qafzeh 11	Homo sapiens	110,000	1,280	1	Holloway et al., 2004
Qafzeh 6	Homo sapiens	110,000	1,568	1	Holloway et al., 2004
Qafzeh 9	Homo sapiens	110,000	1,531	1	Holloway et al., 2004
Ganovce	Homo neanderthalensis	90,000	1,320	1	Holloway et al., 2004
Border Cave	Homo sapiens	90,000	1,522	1	Hughes, 1990
La Ferrassie	Homo neanderthalensis	70,000	1,640	1	Holloway et al., 2004
Spy 1	Homo neanderthalensis	68,000	1,305	1	Holloway et al., 2004
Spy 2	Homo neanderthalensis	68,000	1,553	1	Holloway et al., 2004
Liujiang	Homo sapiens	67,000	1,567	1	Wu et al., 2008
La Quina 5	Homo neanderthalensis	65,000	1,172	1	Holloway et al., 2004
LB1	Homo floresiensis	60,000	419	1	Falk
Shanidar 1	Homo neanderthalensis	60,000	1,600	1	Holloway et al., 2004
Shanidar 5	Homo neanderthalensis	60,000	1,550	1	Holloway et al., 2004
Sambungmacan 1	Homo erectus	55,000	1,035	1	Baba et al., 2003
Sambungmacan 3	Homo erectus	55,000	917	1	Broadfield et al., 2001
Sambungmacan 4	Homo erectus	55,000	1,006	1	Baba et al., 2003
Manot Cave	Homo sapiens	54,700	1,100	1	Hershkovitz et al., 2015
Mt. Circeo	Homo neanderthalensis	52,000	1,360	1	Holloway et al., 2004
Gibraltar	Homo neanderthalensis	50,000	1,200	1	Holloway et al., 2004
La Chapelle	Homo neanderthalensis	50,000	1,625	1	Holloway et al., 2004
Ryonggok 3	Homo sapiens	47,000	1,650	1	Bae et al., 2010
Ryonggok 7	Homo sapiens	47,000	1,450	1	Bae et al., 2010
Amud	Homo neanderthalensis	41,000	1,740	1	Holloway et al., 2004
Le Moustier	Homo neanderthalensis	41,000	1,565	1	Holloway et al., 2004
Neandertal	Homo neanderthalensis	40,000	1,525	1	Holloway et al., 2004
Nazlet Khater 2	Homo sapiens	37,600	1,420	1	Holloway et al., 2004
Hofmeyr	Homo sapiens	36,000	1,580	1	Grine et al., 2010
Chatelperron (les Fees)	Homo sapiens	33,000	1,425	1	Ivanhoe, 2010
Wadjak 1	Homo sapiens	33,000	1,464	1	Coon, 1962
Wadjak 2	Homo sapiens	33,000	1,650	1	many sources in Henn
Mladec 5	Homo sapiens	32,000	1,650	1	Holloway et al., 2004
Jingchuan	Homo sapiens	31,500	1,524	1	HaiJun et al., 2010
Mladec 2	Homo sapiens	31,300	1,390	1	Holloway et al., 2004
Mladec 1	Homo sapiens	31,200	1,540	1	Holloway et al., 2004
Ziyang	Homo sapiens	30,000	1,210	1	Phenice & Sauer, 1977
Zhoukoudian (Upper Cave) 1	Homo sapiens	29,000	1,500	1	Holloway et al., 2004
Zhoukoudian (Upper Cave) 2	Homo sapiens	29,000	1,380	1	Holloway et al., 2004
Zhoukoudian (Upper Cave) 3	Homo sapiens	29,000	1,290	1	Holloway et al., 2004
Cioclovina	Homo sapiens	28,500	1,499	1	Kranioti et al., 2011
Combe Capelle	Homo sapiens	28,000	1,570	1	Holloway et al., 2004
Cro-Magnon 1	Homo sapiens	27,700	1,730	1	Holloway et al., 2004
Cro-Magnon 2	Homo sapiens	27,700	1,402	1	Helmuth, 1999
Cro-Magnon 3	Homo sapiens	27,700	1,590	1	Holloway et al., 2004
Grotte des Infants 4	Homo sapiens	27,500	1,775	1	Holloway et al., 2004
Grotte des Infants 5	Homo sapiens	27,500	1,375	1	Holloway et al., 2004
Grotte des Infants 6	Homo sapiens	27,500	1,580	1	Holloway et al., 2004
Sungir 1	Homo sapiens	27,100	1,464	1	Holloway et al., 2004
Dolni Vestonice 13	Homo sapiens	26,600	1,481	1	Franciscus & Vlcek, 2006
Dolni Vestonice 14	Homo sapiens	26,600	1,538	1	Franciscus & Vlcek, 2006
Dolni Vestonice 15	Homo sapiens	26,600	1,378	1	Franciscus & Vlcek, 2006
Brno I	Homo sapiens	26,000	1,600	1	Holloway et al., 2004
Brno III	Homo sapiens	26,000	1,304	1	Holloway et al., 2004
Dolni Vestonice 3	Homo sapiens	26,000	1,285	1	Franciscus & Vlcek, 2006
Pavlov 1	Homo sapiens	26,000	1,472	1	Franciscus & Vlcek, 2006
Predmosti 10	Homo sapiens	26,000	1,452	1	Holloway et al., 2004
Predmosti 3	Homo sapiens	26,000	1,580	1	Holloway et al., 2004
Predmosti 4	Homo sapiens	26,000	1,250	1	Holloway et al., 2004
Predmosti 9	Homo sapiens	26,000	1,555	1	Holloway et al., 2004
Sungir 3	Homo sapiens	26,000	1,361	1	Holloway et al., 2004
WLH 50	Homo sapiens	26,000	1,540	1	Kubo and Kono, 2011
Dolni Vestonice 16	Homo sapiens	25,600	1,547	1	Franciscus & Vlcek, 2006
Barma Grande II	Homo sapiens	24,800	1,748	1	Parenti, 1973
Sungir 5	Homo sapiens	24,000	1,453	1	Holloway et al., 2004
Veyrier 1	Homo sapiens	24,000	1,430	1	Holloway et al., 2004
Kostenki 2	Homo sapiens	23,800	1,605	1	Holloway et al., 2004
Sungir 2	Homo sapiens	23,800	1,267	1	Holloway et al., 2004
Brno II	Homo sapiens	23,700	1,500	1	Holloway et al., 2004
Arene Candide 1-IP	Homo sapiens	23,400	1,490	1	Ruff et al., 1997
Kostenki 14	Homo sapiens	22,800	1,222	1	Holloway et al., 2004
Mollet III	Homo sapiens	22,300	1,521	1	Soler et al., 2013
Abri Pataud 1	Homo sapiens	22,000	1,380	1	Holloway et al., 2004
Roc de Sers	Homo sapiens	19,200	1,380	1	Ivanhoe, 2010
Roc de Sers	Homo sapiens	19,200	1,485	1	Ivanhoe, 2010
Cohuna	Homo sapiens	19,000	1,447	1	Brown, website
Minatogawa 1	Homo sapiens	18,300	1,335	1	Kubo & Kono, 2011
Minatogawa 2	Homo sapiens	18,300	1,170	1	Holloway et al., 2004
Minatogawa 4	Homo sapiens	18,300	1,170	1	Kubo & Kono, 2011
Le Placard	Homo sapiens	18,000	1,300	1	Ivanhoe, 2010
Rond-du-Barry 8	Homo sapiens	17,100	1,515	1	Ivanhoe, 2010
Collection: Taforalt, Morocco/Algeria (Epipalaeolithic site)	Homo sapiens	16,800	1,572	18	Average of 18 skulls (range 1,279 to 1,738 cc); Ferembach, 1965
Taza, Algeria	Homo sapiens	16,000	1,125	1	Meier et al., 2003
St. Germain-la-Riviere	Homo sapiens	15,800	1,354	1	Movius, 1975
Bruniquel 2	Homo sapiens	15,300	1,555	1	Movius, 1975
Chancelade	Homo sapiens	15,300	1,530	1	Holloway et al., 2004
Cap Blanc 1	Homo sapiens	15,000	1,434	1	Movius, 1975
Mas d’Azil	Homo sapiens	14,000	1,275	1	Movius, 1975
San Teodoro 1	Homo sapiens	14,000	1,565	1	Holloway et al., 2004
San Teodoro 2	Homo sapiens	14,000	1,569	1	Holloway et al., 2004
San Teodoro 3	Homo sapiens	14,000	1,560	1	Holloway et al., 2004
San Teodoro 5	Homo sapiens	14,000	1,484	1	Holloway et al., 2004
Collection: Coobool Creek, Australia (Holocene skeletal collection)	Homo sapiens	14,000	1,386	24	Average of 24 skulls (range 1,117 to 1,541 cc); Brown, website
MLDG, China	Homo sapiens	13,900	1,327	1	Curnoe et al., 2012
Les Hoteaux	Homo sapiens	13,000	1,403	1	Ivanhoe, 2010
Ortucchio 1	Homo sapiens	12,600	1,302	1	Ivanhoe, 2010
Oberkassel 2	Homo sapiens	12,200	1,370	1	Movius, 1975
Keilor	Homo sapiens	12,000	1,497	1	Brown, website
Laugerie-Basse	Homo sapiens	12,000	1,505	1	Ivanhoe, 2010
Arene Candide 1	Homo sapiens	11,800	1,414	1	Ruff et al., 1997
Arene Candide 2	Homo sapiens	11,800	1,424	1	Ruff et al., 1997
Arene Candide 4	Homo sapiens	11,800	1,520	1	Ruff et al., 1997
Arene Candide 5	Homo sapiens	11,800	1,661	1	Ruff et al., 1997
Le Bichon 1	Homo sapiens	11,700	1,465	1	Ivanhoe, 2010
Oberkassel 1	Homo sapiens	11,600	1,500	1	Holloway et al., 2004
Collection: Afalou, Algeria (Epipalaeolithic site)	Homo sapiens	11,500	1,540	25	Average of 25 skulls (range 1,154 to 1,774 cc); Vallois, 1952
Nacurrie	Homo sapiens	11,400	1,597	1	Brown, website
Moča	Homo sapiens	11,300	1,277	1	Sefcakova et al., 2011
Cheix	Homo sapiens	11,000	1,470	1	Ivanhoe, 2010
Duruthy	Homo sapiens	11,000	1,140	1	Movius, 1975
Tuinplaas	Homo sapiens	11,000	1,589	1	Hughes, 1990
Naivasha	Homo sapiens	10,800	1,453	1	Tobias, 1971
Aveline’s Hole, Somerset	Homo sapiens	10,300	1,450	1	Palmer, 1959
Boskop	Homo sapiens	10,000	1,650	1	Tobias, 1971
Mai Da Nuoc	Homo sapiens	10,000	1,430	1	Cuong, 1987
Tepaxpan	Homo sapiens	10,000	1,540	1	Phenice & Sauer, 1977
Hotu 2	Homo sapiens	9,300	1,404	1	Angel, 1952
Hotu 3	Homo sapiens	9,300	1,440	1	Angel, 1952
Culoz I	Homo sapiens	9,000	1,535	1	Ferembach 1974b
Culoz II	Homo sapiens	9,000	1,644	1	Ferembach 1974b
M4300, Bass Strait	Homo sapiens	9,000	1,265	1	Murray et al., 1982
Gamble’s Cave 4	Homo sapiens	8,200	1,470	1	Tobias, 1971
Gamble’s Cave 5	Homo sapiens	8,200	1,530	1	Tobias, 1971, Leakey, 1935
Gramat	Homo sapiens	8,000	1,861	1	Ferembach 1974b
Mai Da Dieu	Homo sapiens	8,000	1,271	1	Cuong, 1986
Teviec 16	Homo sapiens	8,000	1,428	1	Balzeau, PC
Teviec 8	Homo sapiens	8,000	1,430	1	Balzeau, PC
Teviec 9	Homo sapiens	8,000	1,420	1	Balzeau, PC
Rastel	Homo sapiens	7,500	1,682	1	Ferembach 1974b
Song Terus	Homo sapiens	7,500	1,570	1	Balzeau, PC
Fish Hoek 1	Homo sapiens	7,200	1,600	1	Drennen, 1929
Oakhurst 2	Homo sapiens	7,120	1,410	1	Drennan, 1937
Haverty 5, CA	Homo sapiens	7,000	1,380	1	Brooks et al., 1990
Haverty 7, CA	Homo sapiens	7,000	1,310	1	Brooks et al., 1990
Lake Nitchie	Homo sapiens	7,000	1,510	1	Brown, website
Montclus	Homo sapiens	7,000	1,687	1	Ferembach 1974b
Plettenberg Bay	Homo sapiens	7,000	1,350	1	this study
Punin	Homo sapiens	7,000	1,275	1	Phenice & Sauer, 1977
Oakhurst 1	Homo sapiens	6,180	1,403	1	Drennan, 1937
Oakhurst 3	Homo sapiens	5,990	1,195	1	Drennan, 1937
Oakhurst 15	Homo sapiens	5,700	1,214	1	Drennan, 1937
Lansing	Homo sapiens	5,500	1,538	1	Hrdlicka, 1903
Collection: Columnata, Algeria (Epipalaeolithic site)	Homo sapiens	5,500	1,425	3	Average of 3 skulls (range 1,351 to 1,498 cc); Chamla, 1970
Oakhurst 8b	Homo sapiens	5,450	1,247	1	Drennan, 1937
Oakhurst 13	Homo sapiens	5,330	1,290	1	Drennan, 1937
Asselar	Homo sapiens	5,000	1,520	1	Tobias, 1971
Hutton, Somerset	Homo sapiens	5,000	1,415	1	Palmer, 1959
Oakhurst 5	Homo sapiens	4,100	1,389	1	Drennan, 1937
Urtiaga A	Homo sapiens	3,500	1,510	1	Ferembach 1974b
Urtiaga B	Homo sapiens	3,500	1,579	1	Ferembach 1974b
Sauveterre	Homo sapiens	3,400	1,769	1	Ferembach 1974b
Byneskranskop II	Homo sapiens	3,190	1,270	1	this study
Byneskranskop I	Homo sapiens	3,120	1,270	1	this study
Danes’ Graves 1	Homo sapiens	2,250	1,325	1	Wright, 1903
Danes’ Graves 7	Homo sapiens	2,250	1,325	1	Wright, 1903
Elmenteita B	Homo sapiens	2,000	1,585	1	Hughes, 1990
Glastonbury, Somerset	Homo sapiens	2,000	1,468	1	Palmer, 1959
Langwith, Derbyshire	Homo sapiens	2,000	1,250	1	Gladstone, 1921
Loulan	Homo sapiens	1,750	1,265	1	Keith, 1929
Loulan LT 03	Homo sapiens	1,750	1,230	1	Keith, 1929
Niya	Homo sapiens	1,700	1,140	1	Keith, 1929
Broederstroom	Homo sapiens	1,330	1,232	1	De Villiers via Henneberg
Kafue I	Homo sapiens	1,330	1,436	1	Gabel, 1962
Kafue II	Homo sapiens	1,330	1,559	1	Gabel, 1962
Kafue III	Homo sapiens	1,330	1,340	1	Gabel, 1962
Kafue VI	Homo sapiens	1,330	1,362	1	Gabel, 1962
Kafue XIV	Homo sapiens	1,330	1,508	1	Gabel, 1962
Soutpansberg	Homo sapiens	1,310	1,464	1	Steyn and Loots via Henneberg
Collection: Bambandyanalo (K2/Mapungubwe), South Africa	Homo sapiens	1,060	1,408	7	Average of 7 skulls (range 1,280 to 1,565 cc); Galloway, 1959
Gold Mines (Zimbabwe)	Homo sapiens	1,020	1,161	1	Keith, 1924 via Henneberg
Mupungubue	Homo sapiens	1,000	1,591	1	Galloway, 1937
Dias Negress	Homo sapiens	990	1,535	1	Winkler & Holler, 1988
Bristol 1	Homo sapiens	750	1,526	1	Bedde, 1907
Bristol 10	Homo sapiens	750	1,688	1	Bedde, 1907
Bristol 11	Homo sapiens	750	1,655	1	Bedde, 1907
Bristol 2	Homo sapiens	750	1,859	1	Bedde, 1907
Bristol 3	Homo sapiens	750	1,381	1	Bedde, 1907
Bristol 4	Homo sapiens	750	1,608	1	Bedde, 1907
Bristol 7	Homo sapiens	750	1,390	1	Bedde, 1907
Bristol 9	Homo sapiens	750	1,422	1	Bedde, 1907
Collection: Pecos Pueblo, New Mexico (archaeological site)	Homo sapiens	500	1,308	29	Average of 29 skulls (range 1,030 to 1,550 cc); Ruff et al., 1997
Paderborne	Homo sapiens	250	1,541	1	Ruff et al., 1997
Cape Flats	Homo sapiens	150	1,230	1	Drennen, 1929
Collection: China (museum and excavation samples)	Homo sapiens	100	1,445	54	Average of 54 skulls (range 1,188 to 1,776 cc); Brown
Collection: India (anatomical samples, Manjunath 2002)	Homo sapiens	100	1,181	50	Average of 50 skulls (range 975 to 1,425 cc); Manjunath
Collection: Murray River valley, Australia (Holocene skeletal collection)	Homo sapiens	100	1,215	91	Average of 91 skulls (range 996 to 1,432 cc); Brown
Collection: Northern China (museum samples)	Homo sapiens	100	1,370	37	Average of 37 skulls (range 1,140 to 1,600 cc); Brown
Collection: Spitalfields, London (crypt burials)	Homo sapiens	100	1,294	39	Average of 39 skulls (range 1,005 to 1,565 cc); Brown
Collection: Swanport, Murray River, Australia (Holocene skeletal collection)	Homo sapiens	100	1,209	50	Average of 50 skulls (range 1,020 to 1,535 cc); Brown
Collection: Terry anatomical collection (United States)	Homo sapiens	100	1,373	94	Average of 94 skulls (range 1,046 to 1,724 cc); Copes

17 How we built the dataset

1. 01

   Download three open files

   Du et al. 2018, 94 fossil crania (the Dryad file needs a login, so we read the identical copy on figshare) ^[2]; DeSilva et al. 2021, 987 rows ^[3]; and their 2023 revised file, 837 rows ^[5]. Chimpanzees: ^[11].

2. 02

   Match skulls

   We matched rows by species and name, with a short list of known renamings, so a skull in two files is one row: 675 of the 2021 and 2023 rows are the same skull, and 80 of Du et al.’s 94 are in the DeSilva files. Ngandong skull numbers differ between papers and are matched only on an equal first value.

3. 03

   Choose one value

   The analysis uses the 2023 value where the skull is in that file, otherwise the 2021 value, otherwise Du et al.’s mean. Every source value stays in the CSV.

4. 04

   Reproduce before adding

   Our 2023-file periods give the authors’ Table S2 counts and averages exactly for all nine epochs, and the 2021 file gives their counts of 578 skulls in the last 100 years and 1,297 cc. We added no skull of our own.

5. 05

   Calculate and label

   Means, quartiles, Welch intervals and the bootstrap fit are scripted (data/analyse.py) and labelled “IQ Metrics calculation”. Every number taken from a paper’s text is checked against the cached text by data/verify_quotes.py (95 checks).