Craniofacial Anthropometry of some Julio-Claudian Portraits
(Summary of initial investigation)

William Storage and Laura Maish 
Email us about this page

Home

Roman Imperial Portraits

Addendum - May 5, 2009 - Since posting this piece in late 2007, we have refined our selection of craniofacial landmarks used for comparing Julio-Claudian portraits, finding that we actually can differentiate many of the portrait types based solely on craniofacial proportion measurements. After our original investigation, we made 3-D scans of some key portraits. The ability to superimpose the resulting computer models allowed us to identify a number of cranial proportions not used in our first look, but very useful in differentiating the portraits. Many of these proportions are more basic and fundamental to an anthropometric description of a head than those listed below. In hindsight, we realized that several such measurements, while useful for characterizing a skull, are of little use in plastic surgery and were therefore not present in the list used by Farkas, et. al. As an example of a key proportion in our new list, consider face width to tragion-nasion distance. This proportion is not useful in aesthetic plastic surgery because ears cannot be moved. It isn't really useful in reconstructive surgery because other features would control the location of the nasal bridge and surrounding structure. The proportion is always much greater on Caligula portraits than on Augustus portraits not recarved from portraits of Caligula. Consequently, it provides a strong clue as to whether a portrait with Augustus's characteristic hair locks - obviously intending to portray Augustus - was derived from an earlier portrait of Caligula.

We'll update this page in the near future.


Hairstyle has long been the principle means of identifying the subjects of Julio-Claudian and later imperial portraits. Recently, a few scholars have proposed that we may be overly reliant on this approach, the Lockenzahlmethode, as named by the Romische Herrscherbild series of books. In order to investigate the possibility that physiognomic characteristics might be undervalued in the process of identifying uncertain portraits, we conducted a preliminary study, collecting measurements of marble portraits, normalized by calculating ratios of various measurements. This technique is in common usage in the analysis of human craniofacial anthropometry for reconstructive and aesthetic plastic surgery.

We used fifty eight of the proportion indices identified by Farkas and Munro (1986), selected for the ability of their facial landmarks to be located in photographs of frontal and profile views of the sculptures. A few of these still required approximations because of the presence of hair or at landmarks that in a human would be better located by palpation. We included sculptures with restored noses in cases where enough of the original nose was present to reasonably estimate vertical position of the pronasale and where enough original marble existed to establish nasofrontal and nasolabial angles. Most proportions involving ears were excluded, since few sculptures retain original ears. Proportions involving facial contours were also excluded, since they cannot be accurately derived from photographs. Laser-scanning of the portraits would provide more and better data, using techniques common in both plastic surgery and character animation for movie special effects (Aung 1995, Enciso 2003).

We wrote a computer program that displays frontal and profile photographs on screen and allows us to identify the facial landmarks (figure 1). The program records the coordinates of the facial landmarks (listed in Table 2) in a database, from which facial proportions and statistical calculations are made and reported. This approach, including use of photos, as opposed to measurements of the head itself, is accepted in facial plastic surgery applications, provided that the photos meet certain criteria (e.g. clarity, facial plane alignment, large subject-camera distance compared to head height). We adhered to the criteria identified in guidelines for use of photos in plastic surgery.

Our analysis of 25 portraits of Augustus, Caligula, Claudius and Germanicus indicates that craniofacial anthropometry is of limited use in portrait identification. A comparison of facial proportions of Augustus and Caligula illustrate the problem. Few statistically significant differences in average measurements of craniofacial proportions and facial angles can be seen (Table 3). Generally, the standard deviation for a given proportion of one subject is comparatively large when compared to the difference between the mean  value of that proportion for the other subjects. Excluding the proposed Caligula/Augustus recarvings does not help the problem at all. Surprisingly, an analysis of the difference between Augustus heads and proposed Caligula/Augustus heads shows that for most of the proportions where a difference exists, it is in the wrong direction. For example, the average mandibulo-upper face height index is lower for Caligula/Augustus heads than for Augustus heads, yet the index is higher for Caligula heads than for original Augustus heads. Similarly, the mean mandibulo-lower face height index of Caligula/Augustus heads is lower than that of original Augustus heads, yet the index is higher for Caligula than for Augustus. Of the 58 proportions analyzed, the 21 listed below were found to be most significant. These proportions are listed below (Table 1).

proportion index proportion name numerator denominator
2 lower face-face height index subnasale – gnathion nasion – gnathion
3 mandibulo-face height index stomion – gnathion nasion – gnathion
4 mandibulo-upper face height index stomion – gnathion nasion – stomion
5 mandibulo-lower face height index stomion – gnathion subnasale – gnathion
8 nasal index alare (r) – alare(l) nasion – subnasale
9 upper lip height-mouth width index subnasale – stomion chelion (r) – chelion (l)
10 cutaneous-total upper lip height index subnasale – labiale superius subnasale – stomion
11 vermilion-total upper lip height index labiale superius – stomion subnasale – stomion
12 vermilion-cutaneous upper lip height index labiale superius – stomion subnasale – labiale superius
14 vermilion height index labiale superius – stomion stomion – labiale inferius
15 chin-mandible height index sublabiale – gnathion stomion – gnathion
19  nose-mouth width index  alare (r) – alare (l) chelion (r) – chelion (l)
23  lower lip-face height index stomion – sublabiale subnasale – gnathion
25 lower lip-chin height index stomion – sublabiale sublabiale – gnathion
28 mandible Width - face height index stomion – gnathion nasion – gnathion
29 mandibular index stomion – gnathion gonion (r)  gonion (l)
30 Face height  index nasion – gnathion trichion –  gnathion
32 mandible width - total face height  index gonion (r) gonion (l) trichion –  gnathion
35 eye fissure  index palpebrale superius palpebrale inferius exocanthion - endocanthion
52 intercanthal - skull base width index  endocanthion (r) – endocanthion (l) tragion (r) – tragion (l)
54 intercanthal - nasal width index endocanthion (r) – endocanthion (l) alare (r) – alare (l)

Table 1: Most significant facial proportions as determined by statistical analysis of 25 Julio-Claudian portraits examined.

 

 
Figure 1. The Getty Caligula/Augustus showing facial landmarks identified in frontal and profile photos
  
  

 v    vertex    highest point of head
 g    glabella    most prominent point between the eyebrows  
 op    opisthocranion    situated in the occipital region of the head is most distant from the glabella  
 eu    eurion    most prominent lateral point on each side of the skull in the area of the parietal and temporal bones  
 tr    trichion    point on the hairline in the midline of the forehead  
 zy    zygion    most lateral point of each of the zygomatic  
 go    gonion    most lateral point on the mandibural  
 sl    sublabiale    Determines the lower border of the lower lip or the upper border of the chin  
 pg    pogonion    most anterior midpoint of the chin, located on the skin surface in the front of the identical bony landmark of the mandible  
 gn    menton (or gnathion)    lowest median landmark on the lower border of the mandible  
 en    endocanthion    point at the inner commissure of the eye fissure  
 ex    exocanthion (or ectocanthion)    point at the outer commissure of the eye fissure  
 p    center point of pupil    Is determined when the head is in the rest position and the eye is looking straight forward  
 or    orbitale    lowest point on the lower margin of each orbit  
 ps    palpebrale superius    highest point in the midportion of the free margin of each upper eyelid  
 n    nasion   intersection of the frontal and two nasal bones
 pi    palpebrale inferius    lowest point in the midportion of the free margin of each lower eyelid  
 os    orbitale superius    highest point on the lower border of the eyebrow  
 sci    superciliare    highest point on the upper border in the midportion of each eyebrow  
 se    sellion (or subnasion)    Is the deepest landmark located on the bottom of the nasofrontal angle  
 al    alare    most lateral point on each alar contour  
 prn    pronasale    most protruded point of the apex nasi  
 sn    subnasale    midpoint of the angle at the columella base where the lower border of the nasal septum and the surface of the upper lip meet  
 sbal    subalare    point at the lower limit of each alar base, where the alar base disappears into the skin of the upper lip  
 ac    alar curvature (alar crest) point    most lateral point in the curved base line of each ala  
 ls    labiale (or labrale) superius    midpoint of the upper vermillion line  
 li    labiale (or labrale) inferius    midpoint of the lower vermillion line  
 ch    cheilion    point located at each labial commissure
 sto    stomion    Intersection of vertical facial midline and horizontal labial fissure between closed lips 
 sa    superaurale    highest point of the free margin of the auricle  
 sba    subaurale    lowest point of the free margin of the ear lobe  
 pa    postaurale    most posterior point on the free margin of the ear  
 obi    otobasion infrious    point of attachment of the ear lobe to the cheek  
 po    porion (soft)    highest point of the upper margin of the cutaneous auditory meatus  
 t    tragion    notch on the upper margin of the tragus  

Table 2: Description of facial landmarks.


 

Prop. No./ Augustus   Caligula   Claudius   Germanicus
Sample count 7  σ 4  σ 8  σ 6  σ
2 0.46 0.01 0.48 0.03 0.46 0.01 0.48 0.01
3 0.33 0.02 0.35 0.02 0.32 0.02 0.34 0.01
4 0.49 0.04 0.54 0.05 0.47 0.03 0.51 0.03
5 0.70 0.03 0.74 0.02 0.7 0.03 0.71 0.03
8 0.55 0.04 0.61 0.03 0.52 0.04 0.56 0.03
9 0.37 0.02 0.34 0.04 0.39 0.04 0.43 0.07
10 0.65 0.06 0.68 0.03 0.59 0.10 0.69 0.04
11 0.35 0.06 0.32 0.03 0.41 0.10 0.31 0.04
12 0.54 0.14 0.48 0.06 0.74 0.35 0.46 0.09
14 0.78 0.17 0.87 0.05 0.79 0.17 0.93 0.07
15 0.65 0.04 0.69 0.03 0.65 0.03 0.71 0.04
19 0.79 0.08 0.85 0.08 0.81 0.08 0.88 0.01
23 0.25 0.03 0.23 0.03 0.24 0.02 0.20 0.03
25 0.55 0.10 0.45 0.07 0.55 0.08 0.41 0.08
28 1.01 0.04 0.96 0.05 0.96 0.02 0.98 0.03
29 0.32 0.01 0.37 0.04 0.33 0.02 0.35 0.02
30 0.41 0.39 0.71 0.02 0.75 0.02 0.79 0.03
32 0.41 0.39 0.67 0.02 0.72 0.02 0.78 0.05
35 0.54 0.07 0.43 0.08 0.43 0.04 0.47 0.04
52 0.21 0.09 0.24 0.03 0.23 0.02 0.24 0.02
54 1.02 0.07 0.91 0.14 0.98 0.09 0.96 0.09

Table 3: Craniofacial proportion values and standard deviations for selected indices.


Some, but not all, of the most significant differences between facial proportions of the portraits are readily seen by most viewers. Compared to the others, Claudius has a relatively large distance from upper lip to bottom of nose and a more triangular-shaped face. These features are often mentioned in art-history pieces on identifications of imperial portraits. Our measurements occasionally - although barely, and not truly statistically significantly - show counterintuitive results. For example, Augustus's mandibulo-lower face height index and related ratios are almost always smaller than those of Caligula, who is typically described using terms like "weak chin". Values of this index for all four subjects fall within one standard deviation of the mean value of this index for humans (all races) as reported by Farkas (1986). In all other measurements except for eye size, we found the portraits to be lifelike.

Art history literature often mentions the supposed application of the canon of Polycleitos to portraits of Augustus. If such a canon in fact existed, it seems unlikely that it included facial proportions, assuming that sculptors of imperial portraits were proficient enough to apply a canon, had it existed. Based on craniofacial measurements that can be made from frontal and profile photos, Augustus is almost indistinguishable from Germanicus. Some skeptical art historians (e.g., Andrew Stewart, 1978) appear to doubt that the canon of Polycleitos amounted to much more than theory. Facial canons formed by artists of the Renaissance were found by Farkas (1986, 2005) and Edler (2006) to be inconsistent with average proportions of humans - and inconsistent with proportions found in faces judged beautiful by surgeons, international beauty contest judges, and popular vote. In summary, it appears that facial canons are little more than numerology.

Summary
Our findings are, in many ways, "negative results". Craniofacial proportions do not so far appear to be useful in differentiating Julio-Claudian portraits. Further analysis might find them to be more useful in determining whether two portraits of less naturalistically rendered portraits were carved by the same hand, or whether two portraits with facial characteristics farther from the mean are likely to be of the same subject. It may also be that the ears and the nose in particular are far too important to not be considered, as they are generally not considered here.

Future Study
We discovered, unexpectedly, that many of the portraits showed certain facial asymmetries not found in normal human faces. Most predominant was a large difference between the vertical distances from left exocanthion to left cheilion compared to the same measurement on the right side of the face. This is somewhat visible in the photo of Claudius below (2nd from left). A nominally horizontal line extended from the corners of his mouth would intersect a line drawn through his pupils at a surprisingly steep angle. We later found that at least one art historian has reported on this condition in portraits, and speculated on the reason for it. We think this warrants another look. The existing tools we developed  to analyze facial proportions will be useful in collecting and reducing data on facial asymmetries. Measurements of facial contours (requiring 3-dimensional scanning) may also show be more informative than the data that can be extracted from photos.

- Bill Storage - 11/11/2007

  


S. Aung, R. Ngim, and S. Lee, "Evaluation of the laser scanner as a surface measuring tool and its accuracy compared with
direct facial anthropometric measurements," British Journal of Plastic Surgery, 48, 551-558, 1995.

(e.g.,) Dietrich Boshung, Die Bildnisse des Augustus, Das romishe Herrscherbild, pt. 1, vol. 2 Berlin: Gebruder Mann Verlag, 1993. DM 290

Raymond Edler , Pragati Agarwal , David Wertheim , and Darrel Greenhill. "The use of anthropometric proportion indices in the measurement of facial attractiveness"
The European Journal of Orthodontics Advance Access published on January 13, 2006, DOI 10.1093/ejo/cji098.

Farkas, L. G., Katic, M. J., et. al., "International anthropometric study of facial morphology in various ethnic groups/races", Journal of Craniofacial Surgery, 16, 615 (2005).

Leslie G.,. Farkas  and Ian R. Munro. Anthropometric Facial Proportions in Medicine, 1986. ISBN-13: 978-0398052614.

Reyes Enciso, Alex Shawa, Ulrich Neumannb, and JamesMaha, 3D head anthropometric analysis, 2003 Online PDF document

Andrew Stewart; A. F. S. "The Canon of Polykleitos: A Question of Evidence," The Journal of Hellenic Studies, Vol. 98. (1978), pp. 122-131.

 

Copyright 2007 William Storage. All rights reserved.