Human facial parameters, for character work

Each measure is split by sex. The solid band is that sex's typical spread (±1 SD - about 68% of people); the pale outer strip extends to ±2 SD (about 95%), which is the cusp between the second and third standard deviations and a useful outer limit for a character creator. The tick is the mean, the number is the mean value, and both bars share one scale.

Three things break most facial-measurement tables, and they are worth naming up front:

• Landmark choice. “Nose length” means n–sn (nasion to subnasale, the vertical axis) to one author and n–prn (to the tip, a diagonal) to another. The two differ by 5–10 mm on the same nose. Every table here names its landmarks.
• Population. Adult means shift several millimetres between groups — more between groups than between the sexes in some cases. Where a figure is single-population it says so. Published reference ranges (“norms”) are tied to one population each and go stale; treat them as a frame, not a target.^[7]
• Living subject vs dry bone. Skeletal studies (dry skulls, CBCT) report on bare bone; anthropometric studies measure skin over it. Soft tissue adds roughly the same 3–6 mm everywhere, but not uniformly.

Sex is treated here as the binary the source studies used (male/female). The honest summary is that region-by-region dimorphism is larger and more reliable than any whole-face measure — the robust differences are in the jaw, brow and nose, and the whole-face ratios (like fWHR) are the weakest signals of the lot.

Landmarks: g glabella · op opisthocranion (back of skull) · eu euryon (widest point of the vault) · v vertex (top)

Everything else on this page hangs off the braincase. The skull is effectively two modules welded together: the neurocranium (the vault that houses the brain) and the viscerocranium (the face). They grow on different schedules and are modular to a real degree — a large vault with a modest face, and the reverse, both occur naturally. Head shape is captured by a single ratio, the cephalic index.

1.1 Head shape — the cephalic index

Head breadth ÷ head length × 100 sorts skulls into three classic forms. It is the most useful single knob for how a face reads, because the vault underneath sets the outer limit the face has to fit inside.

Distribution is population-specific and not a fixed marker: one 600-skull Indian study found dolichocephaly dominant (47.2%), a Tharu-community study found dolichocephaly in 44.4% of males, and Korean children were mostly mesocephalic (43.2%).^[31][32] Boas showed head shape shifting measurably within one generation of migration, so treat it as a soft prior, never a hard category.

1.2 Sex and size in the vault

• Males are larger in essentially every cranial region, and the most dimorphic spots are the glabella and mastoid process — the same brow region §7 covers.^[16]
• Brain volume averages roughly 10–15% larger in males, and intracranial space about 12% larger — a difference that persists after adjusting for body height.^[27][28]
• Head circumference is a usable proxy for brain volume in this context: it correlates with measured brain volume (ρ = 0.47) and with height (ρ = 0.39), and differs reliably between the sexes.^[26]
• Bone thickness runs the other way. Females carry greater cranial vault cortical thickness than males, and a greater age-related increase in it.^[40]

1.3 The vault keeps changing with age

The adult skull is not finished. Geometric-morphometric work on head CT across ages 20–100 found significant shape change with age in almost every region in males (all but the posterior cranial fossa), and in females mainly in the anterior and middle cranial fossae.^[39] The pattern is localised compression of the inner frontal and posterior parietal regions with relative expansion of the inferior parietal and temporal regions — which loosely tracks where the brain's grey matter thins.

Practical read: the brow and forehead stay put or thicken, while the midface and orbital rim resorb around them (§9). That contrast is a large part of why an older face looks like the face slid down inside a stable skull.

1.4 Genetics of the vault

Cranial vault shape is highly heritable and overwhelmingly polygenic, like the face — but with a different lead actor.

• RUNX2 is the master regulator of calvarial (flat-bone) ossification; it sits at the centre of a network of FGF, BMP/TGF-β, Wnt, Hedgehog and PTHLH signals.^[34]
• A multi-ancestry GWAS of 3D vault shape (6,772 children) found 30 genome-wide loci explaining ≈1.31% of global vault shape variation; a pleiotropy-informed re-analysis raised that to 120 loci.^[33][34]
• ≈60% of vault loci overlap facial loci, and the frontal region shares signal with both face and brain — which is exactly why head shape and face shape are not independent variables.^[33]
• Cephalic index specifically has been associated with loci near SOX9 and SOX11. The old textbook candidate FGFR1 did not replicate.^[33]

Landmarks: n nasion · gn gnathion · zy zygion · tr trichion (hairline) · en endocanthion · ex ectocanthion · al alare · go gonion · st stomion

The face measures in §3–§8 only mean something against the head that contains them. Two families of ratio do the connecting work: the facial index (how long the face is for its width) and the face-to-head width ratios (how much of the skull's width the face actually fills).

2.1 Facial index — face length against face width

Facial index = (nasion → gnathion) ÷ (zygion → zygion) × 100. It is the standard classification of overall face form.

Adult means in one 200-student Indian sample: males 88.15 ± 6.88, females 85.74 ± 5.26, full range 75.9–110.0 — i.e. the sexes straddle the meso/lepto boundary and individuals spread across the entire range.^[36] A parallel upper facial index divides upper-face height by bizygomatic width instead, banding from hypereuryne (<44.9) up to leptene (>54.9).

2.2 Face width against head width

These are the ratios that actually change a silhouette. They are stated as fractions of each other rather than absolutes, so they survive scale.

2.3 The neoclassical canons — useful scaffolding, poor law

A set of ratios handed down from classical art, still used as a first-pass scaffold. Stated as equations because that is how they are applied:

A systematic review of raw data found canons 2, 6, 7 and 8 differ significantly between ethnic groups, and the greatest variability overall sits in forehead height, intraocular distance and nasal width — three of the canons' core assumptions. Use them as a starting grid to deviate from, not a target.

2.4 How head shape bends the face

• Brachycephalic head (short, broad): the vault is wide, so bizygomatic sits near head width and the face reads broad and flatter; the midface tends to look deeper-set because there is more transverse space around it.
• Dolichocephalic head (long, narrow): the vault is narrow, so the same cheekbone width occupies a smaller fraction of it; the face reads longer and slimmer, and features crowd closer to the midline.
• Because bizygomatic tracks head breadth at roughly 88–95%, widening the skull automatically widens the face. A narrow face on a wide head reads pinched; a wide face on a narrow head reads as the cheekbones dominating.
• Vertical coupling is looser than horizontal. The face can be long or short on the same head breadth — which is exactly what the facial index in §2.1 measures, independently of the cephalic index in §1.1. Head shape sets width; facial index sets length.

Landmarks: zy zygion (most lateral point of the zygomatic arch) · fmo frontomalare orbitale

The zygomatic bones set the widest point of the face (bizygomatic width) and define the malar “apple” under the eye. They are the single most dimorphic width measure in the face when body size is controlled for — a much stronger signal than the width-to-height ratio it feeds into (§8).

What actually varies

• Overall breadth — the main dial; male arches sit further laterally, giving the wider, flatter midface.
• Anteroposterior (AP) projection — how far the cheek stands off the face in profile. “Primary determinant of the malar apple.”
• Arch orientation — more front-facing vs more lateral sweep; the classic trait differences in how “flat” the face reads front-on.
• Muscle over bone — masseter and temporalis bulk thicken the visible cheek independently of the bone, and scale with body mass.

Population spread is real and large: in a 3,000-subject Chinese survey, face width was 147.6 mm in men and 140.1 mm in women, with jaw width 118.5 / 114.5 and face length 117.0 / 109.7 — several millimetres wider than the North-American figures above.^[41]

Landmarks: go gonion (jaw corner) · cdl condylion (top of the joint) · gn gnathion (chin)

Four numbers carry almost all of the jaw's shape information: the gonial angle (how square vs rounded the corner is), the bigonial width (jaw width at the corners), the bicondylar breadth (width across the joints) and the ramus height (vertical length of the back of the jaw). Bigonial width is widely used as one of the best single-bone sex discriminators.

Sex differences — the reliable part

• Males: higher ramus, more pronounced gonial angle (sharper, everted corner from heavier masseter attachment), larger inter-gonial width and a more distinct antegonial notch. Mandible-based sexing from 3D shape reaches ~91% accuracy.^[4]
• Females: more obtuse (larger) mandibular angle, relatively wider ramus, and a more gracile body — a rounder, softer lower face.^[4]
• Magnitude: the inter-gonial width difference lands around 7.6% in a well-matched sample.^[4]

Age & dentition

The mandibular angle increases by roughly 3–7° from youth to old age, and alveolar bone resorbs after tooth loss, shortening the ramus and reducing chin projection.^[8] A square young jaw reads increasingly rounded and receding with age.

Landmarks: n nasion (bridge root) · prn pronasale (tip) · sn subnasale (base of columella) · al alare (widest wing) · ac alar crease

5.1 Size — and why the numbers disagree

The nose has the most population spread of any feature, and also the most measurement confusion. Two families of number float around: n–sn / n–prn length (vertical and diagonal length) and alar width. Reported means for “nose length” range from ~41 mm to ~57 mm across studies largely because of landmark choice as much as population:

Sources: cross-study comparison in the Indian rhinoplasty anthropometry paper^[10] and the Indonesian 3D reference set.^[11]

5.2 Width & the nasal index

The nasal index = (alar width ÷ nasal height) × 100 classifies nose form better than absolute width, because it normalises for face size:

Width reference values: white women 31.4 ± 2.0 · Korean-American 35.5 ± 3.4 · Chinese 39.2 · Japanese 36.3 · Afro-American 43.5 · African 45.9 mm.^[10][11]

5.3 Projection — angles & ratios

Profile aesthetics are governed by a small set of angles and ratios. These are the ones with broadly agreed ideals:

Sex & population in the profile

• Size: male noses run larger on average — around 4 mm more on the alar-to-tip distances in one large study.^[14]
• Tip direction: female tips tend upturned; male tips sag/droop slightly, with a straighter, higher-radix dorsum.^[9]
• Projection by group: Caucasian and Middle-Eastern noses carry the most nasolabial tip projection; African-American noses read under-projected on nasolabial lines but ideal against nasofacial ones — the contrast is as much about lip and midface projection as the nose itself.^[13]

Landmarks: p pupil centre · en endocanthion (inner corner) · ex ectocanthion (outer corner)

Interpupillary distance (IPD) sets eye spacing and scales with the orbit and face. It grows through childhood and early adulthood, then plateaus and shrinks slightly in old age.

Related spacing measures

• Intercanthal distance (inner-corner to inner-corner) ≈ 29.5–30 mm (F ≈ 29.5, M ≈ 30.0). The alar base width ideally equals this.
• Canthal tilt — the outer corner sitting above the inner (positive) reads as feminine and youthful; level or negative reads masculine.
• Orbital shape: females trend toward a rounder orbital aperture, males toward a squarer, deeper-set rim.^[16]

Landmarks: g glabella (brow bridge) · so supraorbital rim · bregma top-front of skull

The upper third is one of the strongest perceived-gender regions: observers read masculinity from glabellar and supraorbital prominence, frontal-bone thickness and orbital shape. On average, males project at the glabella and slope away above it; females have a flatter, more vertical forehead with a smaller brow ridge.

Brow-ridge volume is itself measurably dimorphic and is one of the traits forensic anthropologists score for sex; it also varies between populations.^[17]

fWHR = bizygomatic width ÷ upper-facial height (upper lip to brow)

A dimensionless proxy for how “wide and square” a face reads. It is popular and heavily studied, but it is a weak dimorphism signal — and the reason matters for anyone using it as a masculinity dial.

Ageing is not a uniform deflation or a uniform droop. It is site-specific bone resorption under a set of independently-deflating fat compartments, and the two interact.

9.1 The skeleton

Site-specific resorption is heaviest in the orbit, piriform aperture and maxilla; postmenopausal changes are accelerated in women in the orbit, while men show more mandibular remodelling.^[18]

9.2 The fat compartments

Facial fat is partitioned by septal walls into discrete compartments, not spread as a uniform pad — the core insight from Rohrich & Pessa’s cadaver work.^[19] Ageing is compartment-specific deflation, so each pocket should be modelled as its own volume that shrinks and repositions independently.

Face shape is highly heritable but extremely polygenic, and even the largest studies explain only a single-digit percentage of the variation. The practical consequence: facial variation is a broad continuum to be blended, not a set of gene switches to toggle.

Gene → trait map

EDAR V370A is the standout: the derived allele — common in East Asians and Native Americans, absent in Europeans and Africans — reduces chin protrusion and shortens the mandible, while also changing hair thickness, incisor shovelling and sweat-gland density. One variant, many features.^[35]

Named loci worth knowing

• POU3F3 (2q12.1) — strong nose-region signal in the 2025 study.
• PAX3 — nasion position / interorbital septum.^[21]
• PRDM16 — nose width & height; SOX9 — nose shape; EDAR — East-Asian hair/nose traits.^[21][22]
• TP63, C5orf50, COL17A1 — midface and eye-spacing measures.^[21]
• HOXD cluster, BMP4 — overall proportions and jaw width.

The points every measurement above is built from. Naming them is what makes two tables comparable.

Every figure above is attributed. Numbered citations in the text resolve to these.

1. 01Caton & Dixson (2022). Bizygomatic width is highly sexually dimorphic when adjusting for allometry. Biology Letters. pmc.ncbi.nlm.nih.gov/articles/PMC9554718/
2. 02Sex Determination of Human Mandible Using Metrical Parameters (South Indian sample, n=250). pmc.ncbi.nlm.nih.gov/articles/PMC3919368
3. 03Farkas adult craniofacial norms (gonial width M 97 ± 5.8 / F 91 ± 5.9 mm), as tabulated in Ogodescu et al. (2021), Updating Standards of Facial Growth. pmc.ncbi.nlm.nih.gov/articles/PMC8156684
4. 04Chalazoniti, Lattanzi & Halazonetis (2024). Shape variation and sex differences of the adult human mandible. Sci Reports. www.nature.com/articles/s41598-024-57617-7
5. 05Gonial Angle in Forensic Anthropology to Determine Age and Sex. pmc.ncbi.nlm.nih.gov/articles/PMC11288476/
6. 06Gender Determination Using Mandibular Angle Measurement (mandibular angle M 122.19 ± 3.89 vs F 124.72 ± 3.24). brieflands.com/journals/amhsr/articles/114608
7. 07Facial Type — population norms (Plastic Surgery Key). plasticsurgerykey.com/facial-type-3/
8. 08Shaw & Kahn / Pessa: Changes in the Facial Skeleton With Aging. pmc.ncbi.nlm.nih.gov/articles/PMC3404279/
9. 09Using data-driven phenotyping to investigate the impact of sex on 3D human facial surface morphology. pmc.ncbi.nlm.nih.gov/articles/PMC10335371
10. 10Do Aesthetic Average Nasal Parameters Matter For Rhinoplasty in India? (cross-study nasal index & width comparisons). pmc.ncbi.nlm.nih.gov/articles/PMC6848649/
11. 11Anthropometric analysis of the external nose of Indonesian females (Rhinobase 3D reference values). exa.ai/library/publication/vfjw4q7zt1p
12. 12The Ideal Nose — nasal projection, Goode ratio, nasofacial/nasomental angles (Rhinoplasty Archive). www.rhinoplastyarchive.com/articles/rhinoplasty-fundamentals/the-ideal-nose
13. 13Role of AI in Determining Ideal Nasal Tip Projection in Diverse Populations. pmc.ncbi.nlm.nih.gov/articles/PMC11581758/
14. 14Shaffer et al. (2016). A GWAS Identifies Five Loci Influencing Facial Morphology. (male noses ~4 mm larger; high heritability, many small-effect variants). pmc.ncbi.nlm.nih.gov/articles/PMC3441666/
15. 15Evaluation of interpupillary distance in the Turkish population (n≈756, by age/sex). pmc.ncbi.nlm.nih.gov/articles/PMC4529256/
16. 16Quantitative Analysis of Male Versus Female Frontal Bone and Orbital Skeletal Morphology. pmc.ncbi.nlm.nih.gov/articles/PMC12459598
17. 17Shearer et al. (2012). Sexual dimorphism in human browridge volume (3D dry crania). pubmed.ncbi.nlm.nih.gov/22776689
18. 18Facial bone aging: an update and literature review (orbital enlargement, maxillary height loss, mandibular angle increase). www.sciencedirect.com/science/article/pii/S2352587826000185
19. 19Rohrich & Pessa (2007). The Fat Compartments of the Face. Plast Reconstr Surg. pubmed.ncbi.nlm.nih.gov/17519724
20. 20Combined GWAS of facial traits in Europeans (2025). Nature Communications. link.springer.com/article/10.1038/s41467-025-61761-7
21. 21Shaffer et al. / PAX3, PRDM16, TP63, C5orf50, COL17A1 (PLoS Genetics 2016). journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1006149
22. 22Identification of five novel genetic loci related to facial morphology (SOX9, WDR35, HOXD1-MTX2). BMC Genomics 2018. link.springer.com/article/10.1186/s12864-018-4865-9
23. 23Head length norms by population (Hong Kong / British / US / Japanese). Online Anthropometry reference. personal.cityu.edu.hk/meachan/online%20anthropometry/chapter2/Ch2-26.htm
24. 24Cranial anthropometry in 600 North Indian adults (head breadth M 139.5 / F 136.19 mm). www.ijmhr.org/ijar_articles_vol1_02/231.pdf
25. 25Anthropometric data: head breadth percentiles (Roy Mech). roymech.org/Useful_Tables/Human/Human_sizes
26. 26Microcephaly measurement in adults (n=2,508; head circumference M 56.3 ± 2.6 / F 54.2 ± 2.4 cm; correlation with brain volume and height). pmc.ncbi.nlm.nih.gov/articles/PMC9126574/
27. 27Are sex differences in human brain structure associated with sex differences in behavior? (male brains 10–15% larger). pmc.ncbi.nlm.nih.gov/articles/PMC8726594/
28. 28Males and females differ in specific brain structures (intracranial volume ~12% larger in males, n>14,000). University of Cambridge. www.cam.ac.uk/research/news/males-and-females-differ-in-specific-brain-structures
29. 29Brain size — endocranial volume (≈1,260 cm³ men / ≈1,130 cm³ women). en.wikipedia.org/wiki/Brain_size
30. 30Cephalic index — dolichocephalic / mesaticephalic / brachycephalic classification. en.wikipedia.org/wiki/Cephalic_index
31. 31Cephalic Index in the Indigenous Tharu Community (dolichocephaly 44.4% of males). pmc.ncbi.nlm.nih.gov/articles/PMC8959361/
32. 32Establishment of Cephalic Index using cranial parameters (dolichocephalic 47.2%). pmc.ncbi.nlm.nih.gov/articles/PMC8254511/
33. 33Joint multi-ancestry and admixed GWAS of human cranial vault shape (30 loci; ≈1.31% of vault shape variance; SOX9/SOX11 for cephalic index; 60% overlap with facial loci). Nature Communications 2023. www.nature.com/articles/s41467-023-43237-8
34. 34Enhanced insights into the genetic architecture of 3D cranial vault shape using pleiotropy-informed GWAS (120 loci; RUNX2). Communications Biology 2025. www.nature.com/articles/s42003-025-07875-6
35. 35A genome-wide association scan implicates DCHS2, RUNX2, GLI3, PAX1 and EDAR in human facial variation. Nature Communications 2016. www.nature.com/articles/ncomms11616
36. 36An Anthropometric Study of the Facial Index (facial index classes; M 88.15 / F 85.74). pmc.ncbi.nlm.nih.gov/articles/PMC13399996
37. 37Cheekbone prominence and midface width ratios (bitemporal/bizygomatic 75–85%; bigonial/bizygomatic 70–82%). facialharmonyai.com/blog/cheekbone-prominence-bizygomatic-ratio
38. 38Historical Tools of Anthropometric Facial Assessment: the neoclassical canons (systematic review). Aesthetic Surgery Journal 2022. academic.oup.com/asj/article/42/1/NP1/6369423
39. 39Evaluation of morphological changes in the adult skull with age (cranial fossae shape change; males across nearly all regions). pmc.ncbi.nlm.nih.gov/articles/PMC5108156/
40. 40Evaluation of skull cortical thickness changes with age and sex (female cortical thinning 36–60%; greater vault thickness in females). JBMR 2015. onlinelibrary.wiley.com/doi/full/10.1002/jbmr.2613
41. 41Measurement and analysis of human head-face dimensions (n=3,000; face width, jaw width, face length, nose protrusion). pubmed.ncbi.nlm.nih.gov/18727867

What changed from the original facial-character-parameters.md:

• Errors fixed. “Romanose profile” → Roman nose. Two broken pseudo-code formulas were removed — mandibularWidth = gonialAngle × 0.87 multiplies a degree value by a unitless factor to produce a millimetre width, which cannot be right.
• Genetics corrected. “253 loci / ~7% / h²=0.226” rewritten to the real figures: 253 SNPs across 188 loci, up to 7.9% per trait (4.5% whole-face), SNP-h² ≈ 0.23.
• Blank cells filled. Every “–” in the original tables now carries a sourced value or an explicit marker.
• IPD corrected. The draft’s population ranges were unsourced and spanned the full adult distribution; replaced with an age/sex reference set.
• Sections added: the cranium (head size, cephalic index, vault sex/age/genetics), how the face sits in the head (facial index, width ratios, the canons), projection angles & ratios, facial width-to-height ratio, brow & forehead, skeletal ageing, and the landmark glossary.
• Head-shape links. Face widths are now tied to head shape: bizygomatic tracks head breadth at ≈88–95%, the bigonial/bizygomatic ratio is called out as the biggest single silhouette driver, and the cephalic index is connected to the facial index (§2.4).
• Vault genetics added. RUNX2 and the calvarial ossification network, 30→120 vault loci, and the ≈60% overlap with facial loci — plus a gene→trait table for PAX3, EDAR, DCHS2, RUNX2, GLI3.
• Key-measure chart split by sex. The summary strips were pooled; they now carry a separate female and male bar per measure, each with its own ±1 SD band and mean tick on a shared scale, so the two are directly comparable. Head breadth shows the 5–95th range instead of ±1 SD (none published).
• ±2 SD tier added. Each bar now carries a pale outer strip running to ±2 SD — the cusp between the second and third standard deviations, ≈the 2nd and 98th percentiles — so the chart shows both the typical spread (~68%) and a defensible "extreme" limit (~95%) for character extremes. Scales were widened to fit. fWHR uses the pooled SD (no per-sex SD published) and head breadth's SD is derived from its published range.
• Kept from the draft: the fat-compartment model, the checkbox “what varies” lists, and the polygenic-blending conclusion — all accurate and useful.
• Sources are now linked and numbered instead of a prose blob.