The Foramen Magnum, Bipedalism, Habitat, and Human Ancestry: A computer study of 3D digital skulls.


The use of the foramen magnum (FM), the opening where the spinal cord enters the brain case, as an indicator of bipedalism has been controversial for years. It was first used in the 1920’s when Raymond Dart argued that Taung child (Australopithecus africanus) was bipedal. The question of whether the FM can be used as an indicator of locomotion is important because if so, it would allow anthropologists to determine the posture of a specimen using only a skull. A specimen's posture is important in determining its place in the human lineage. I hoped to evaluate this question using modern computer tools.

This project compares correlations between different measurements of FM placement and posture. Using 3D scans from the Internet allowed more consistent measurements of specimens than 2D photos, and these scans were more accessible than the skulls themselves. I compared humans, extant primates, and several fossil specimens thought to be human ancestors. I compared a measure similar to Dart's original to one intended to exclude adaptations in muzzle length, and compared angles the FM makes with facial features and the upper jaw.

My results support a relationship between FM placement and bipedalism, but contradict previous conclusions about the Australopiths. However the angle the FM makes to the plane of the upper jaw (maxilla) was a better indicator of habitat, and could be used to evaluate fossil specimens' place in the human lineage. This will help us better understand our origins and place in nature.

Question / Proposal

Is foramen magnum (FM) placement only a function of posture or is it influenced by other adaptations such as eating, and can the FM's position or orientation be used to determine human ancestry?

I will be testing four different measurements. The first three concern the FM's position relative to various landmarks of the face and skull. The fourth measurement compares the FM's orientation to the plane of the upper jaw ("maxilla").

I expect that previous measurements were influenced by the short muzzle found in humans and specimens such as the Taung Child as opposed to other apes, and so measurements which exclude muzzle length will give different results.

I predict that the FM's orientation relative to the maxilla will have a better correlation to posture. I assumed the maxilla indicates the normal rest position of the face because the upper jaw is usually held parallel to the ground while eating. The orientation of the FM shows the angle at which the spine meets the skull. This should be more indicative of posture, because the spine in humans is S-shaped to support bipedal posture, whereas in quadrupedal species the spine is more straight. Therefore I expect that in bipedal species the spine will enter at an angle from the front, while in quadrupeds it will enter from a more rear facing angle.


The first time the foramen magnum (FM), the braincase opening for the spinal cord, was used as an indicator of upright posture was in the 1920's by Raymond Dart in his study of the Taung child skull (Australopithecus africanus). This argument has been used and debated ever since. A recent study of FM placement was done in 2013 (Ruso and Kirk, 2013). Their results were challenged by Ruth et al, 2016, who made the argument that due to the landmarks used Ruso and Kirk's measurements were subject to morphological variations other than FM placement. In response Ruso and Kirk published another study using different measurements (Ruso and Kirk, 2017).

There are several major differences between my project and these previous studies:

This project is done with 3D virtual skull scans which can be measured more accurately and consistently using a computer than the 2D images that the aforementioned studies used. All the skulls used in this project were accurate scans of real skulls obtained from academic sources.

Another part of my research was identifying skulls suitable for my approach. I went through over a hundred, my final set of specimens is 63. Among the skulls that I couldn't use was Dart's Taung Child skull because there was too little of the FM left to define the FM's centroid. Other reasons for excluding skulls included low resolution, vertebrae or lower jaw obscuring important landmarks, and missing portions and landmarks of the skull.

I restricted my comparison to humans, primates, and a few fossil specimens that are potential human ancestors, I consider this to be a better comparison than the larger set of mammals, and marsupials that Ruso, Kirk, and Ruth studied. These previous studies were counting "bipedal" rodents and marsupials as a part of their bipedal dataset. These are not true bipeds, they rest on their tail when standing and use it as a counterbalance to move. These "bipedal" rodents and marsupials have a different posture because of this counterbalance. I would expect these animals to have measurements similar to that of quadrupeds and unlike the more vertical posture and S-shaped spine of true bipeds such as modern humans.

I began with the more historical comparison of the FM's location relative to the tip of the upper jaw (maxilla) and the back of the skull. This was similar to Dart's original measurement.

The variability in muzzle length and braincase size between humans and primates is significant; for this reason, two of my measurements used the zygomatic processes near the rear of the jaw (where the jaw muscles pass up the face) as landmarks for comparison. This was to exclude morphological variations in the jaw itself.

Another of the measurements these more recent studies used was a measurement of the FM's orientation to the plane of the upper jaw or maxilla. I compared this and the other measurements for all 63 human and primate specimens, sorting by species and habitat.

Method / Testing and Redesign

3D digital skull models were collected from various Internet sources. A subset was selected for resolution, completeness, and landmark visibility. Using orthographic projections, each model was rotated and aligned so that the same basal view, parallel to the plane of the upper jaw (maxilla), was used, eliminating parallax error in my measurements. These views were exported as 2D images and used to identify the centroids of the foramen magnum (FM) and zygomatic processes (ZPs) for the following three measurements. By using either ratios between lengths or angles my measurements are independent of scale.

Measurement 1 ("Dart's Ratio", Figure 1) is the ratio of the distance between the FM and the midline front of the maxilla (tip of the jaw) to the distance between the FM and the midline back of the skull. This measurement is similar to the measurement that Dart made on Taung Child. My measurements use the centroid of the FM; Dart's used the basion (the midline point on the front rim of the FM). I used the centroid because the basion's position could vary depending on the length and size of the FM relative to the brain case.

Measurement 2 ("ZP Angle", Figure 2) is an angular measurement of the FM's position relative to the ZPs; it is defined as the angle from one ZP to the other, with the FM as the vertex. In my preliminary, casual survey of greater ape skulls these landmarks appeared to roughly form an equilateral triangle, suggesting a consistent primate geometry unrelated to posture across species. I collected this data to explore ways of removing adaptations unrelated to posture from my measurements.

Measurement 3 ("ZP Ratio", Figure 3) is intended to remove differences in muzzle length from FM measurements to eliminate jaw adaptations from the question of posture. The ZPs encompass the path of jaw muscles up the face. I found the intersection of a line connecting the ZP centroids and the skull's midline. Measurement 3 is the ratio of the distance between the FM and this intersection to the length between the FM and the midline point at the back of the skull.

Measurement 4 ("FM Orientation", Figure 4) shows the orientation of the FM, and was taken from orthographic projections of the side of the skull. The line defined by intersection of the skull's midline (medial plane) with the front and back rim of the FM was compared to the plane of the maxilla. In this way a value of 0 degrees was a downward facing FM, negative values an anterior facing FM, and positive values a dorsal facing FM.

The error margins of each skull for each of the methods were estimated during measuring. All measurements and error margins were entered into a spreadsheet for comparison. Results were plotted by specimen, species, or genus as appropriate, as were comparisons by habitat.


My measurements are presented in the following table (Figure 5). I plotted my results in the whisker bar plots below. These plots are organized by fossil specimen, species within the genus Homo, genus, and also by habitat for apes, Old, and New World monkeys (note that only one species of ape is representative of each habitat, and baboons are the only example used of a terrestrial monkey).

As I stated in my "Research" section I discarded Taung Child as too incomplete, but I included fossil specimens of Australopithecus afarensis and Australopithecus/Paranthropus boisei.

Figures 6-7 are plots of Dart's ratio. Smaller values indicate a more anterior or ventral FM placement; larger values are a more posterior or dorsal FM placement. Humans and our Homo erectus cousins clearly have a more anterior position, while the great apes and various monkeys occupy overlapping positions on this chart.

Figures 8-9 are plots of ZP ratio. This measurement was intended to remove the possible interference of muzzle length in Dart's ratio from the question of FM placement. The plots however, show that other than a tightening of the spread between bipeds and quadrupeds, there isn't a significant difference in the overall pattern of the results. This affirms the use of Dart's ratio as an indicator of posture, contradicting my hypothesis that Dart's ratio was influenced by muzzle adaptations.

Although small values for Dart's ratio and the ZP ratio correlate to bipedal posture, they don't show a reliable correlation to habitat between apes and monkeys (Figures 7, 9). Arboreal New World monkeys, for example, have a much smaller ratio than their Old World counterparts.

Figures 10-11 plot ZP angle. This measurement shows inconsistent correlation to posture and a slight trend that corresponds to habitat. Fossil KNMER406 (Figure 14) illustrates how extreme adaptations affect such measurements. Even excluding KNMER406, there is too much overlap for this measurement to be useful in evaluating individual specimens.

Figures 12-13 The FM orientation provides similar results to Dart's ratio but shows a better correlation to habitat. All the angles for Humans and Homo erectus are at or below -8 degrees, and except for two baboons all the quadrupeds are above this angle. Baboons, the most terrestrial monkeys in my data set, clustered around an FM angle of 0 degrees, or straight down; other primates had a dorsal orientation. The groups with the most dorsal FM orientation are those which hang suspended from their hind feet, as the orangutan, or from prehensile tails, such as New World monkeys. Terrestrial apes (represented by gorillas) and the terrestrial-arboreal apes (chimpanzees) aren't significantly different; this might be attributed to the gorilla's great bulk as opposed to the chimpanzee's agility.

A composite skull of Australopithecus afarensis, AL333, and two skulls of Australopithecus/Paranthropus boisei, KNMER406 and OH5, were measured. KNMER406 was an outlier in the ZP measurement because of its exaggerated zygomatic processes (Figure 14). In every other comparison these specimens fell within the range of Pan and Gorilla, not approaching Homo's FM placement or orientation.


My results confirm that Dart's ratio is a statistically significant indicator of upright posture, as are my ZP ratio and FM orientation measurements. My results show FM orientation is also a reliable indicator of habitat. My results do not support claims that Australopithecus fossil specimens AL333, KNMER 406, and OH 5 were bipedal. Composite skull AL333 (Australopithecus afarensis) overlaps with Pan (chimpanzees) in all the measurements.

My original questions were: Is FM placement only a function of posture or is it influenced by other adaptations such as eating, and can the FM's position or orientation be used to determine human ancestry? My predicted outcome was that Dart's ratio was influenced by muzzle length, and that FM orientation would be a superior indicator of posture. Unlike I predicted, Dart's ratio isn't influenced by other adaptations, since ZP ratio, a measure I formulated to eliminate muzzle length, shows much the same results. But as I predicted, FM orientation was the best indicator of posture, with Homo on one extreme and orangutans which hang from their feet and New Word monkeys with prehensile tails on the other extreme. Overall, Humans and Homo erectus possess significantly different FM placement and orientation from that of other extant and fossil primates.

My approach was mostly successful, but I would discard further efforts to measure the geometry comparing the angles formed by the FM and ZPs because these angles had little predictive value.

Measurements of the ZP ratio have their greatest value in affirming the FM approach, but Dart's ratio has the advantage of a wider spread of results so that it would be more useful discerning individual species. As I have stated, the FM orientation appears more indicative of habitat and behavior. If a specimen were to fall between bipeds and quadrupeds, then cross referencing its position between these different measurements might give a meaningful determination.

More specimens would have improved my results. I did not have any Neanderthals, gibbons or siamang, nor a number of monkeys such as the Barbary Ape. Many skulls were only available as DICOMs, derived from CT scans. I experienced too many technical difficulties converting these to usable 3D models, so I only included five of these. Skulls provided freely by the Smithsonian Institute's Human Fossil Collection were too low resolution to be used. Other skulls were not freely available. More open sharing by those in the anthropology community would improve studies such as mine.

With a larger set of primate specimens, anthropologists would not only be able to use these measurements to reevaluate the fossil specimens they have now, but also evaluate any future finds. Since my results do not support that the Australopiths were bipeds, perhaps anthropologists should focus their search for early bipeds in some new direction. An improved understanding of the human lineage would help us as humans to better understand who we are, and our place in nature.

About me

I'm Nerine Storm Thompson. I enjoy exploring science, painting, and playing the piano. My favorite science subjects are physics and anthropology. Physics is looking ahead, and anthropology is looking behind; I think there's a lot more to be discovered in both. I like to know about the past, but I want to contribute to the future, so I intend to become a physicist.

I'm doing this project because there is a lot of debate about whether certain fossils were bipedal or not, as well as the methods used to decide this. And it occurred to me that foramen magnum placement was something I could measure just using some of the 3D programs on my computer and 3D models that are available on the Internet. I was delighted that I might be able to clear up some of the debate, and by extent help people better understand their origins.

Winning the Google Science Fair would mean a lot to me. Success would tell me that my work is appreciated and would make me more confident about my efforts. As a homeschooler I hope to improve my academic portfolio to help me to get into a university.

Health & Safety

My research was done under the guidance of my parents John and Nanette Thompson (

I did this project at home on my personal desktop computer. This project was a computational project involving data files, image processing and computer computations. I had no personal contact with actual physical specimens, so there were not any special safety procedures applicable to my work.

Bibliography, references, and acknowledgements


Been, E., et al. (2014). "Foramen Magnum Orientation and Its Association with Cervical Lordosis: A Model for Reconstructing Cervical Curvature in Archeological and Extinct Hominin Specimens." Advances in Anthropology. 4:133-140. <>

Cignoni, P., et al. (2008). “MeshLab: an Open-Source Mesh Processing Tool.” Sixth Eurographics Italian Chapter Conference. 129-126. <>

Dart, R.A. (1925). “Australopithecus africanus The Man-Ape of South Africa.” Nature. 115:195-199. <>

Fedorov, A., et al. (2012). "3D Slicer as an Image Computing Platform for the Quantitative Imaging Network." Magnetic Resonance Imaging. 30(9):1323-41. <>

Killgrove, K. (2015). "How To Print Your Own 3D Replicas Of Homo Naledi And Other Hominin Fossils." Forbes. <>

Kimbel, W.H., White, T.D., and D.C. Johanson (1984). “Cranial morphology of Australopithicus afarensis: a comparative study based on a composite reconstruction of the adult skull.” American Journal of Physical Anthropology. 64(4):337-388. <>

Kimbel, W.H. and Y. Rak (2010). "The cranial base of Australopithecus afarensis: new insights from the female skull." Philosophical Transactions of the Royal Society B. 365:3365–3376. <>

Leakey, L.S.B. (1959). “A New Fossil Skull From Olduvai.” Nature. 184:491-493. <>

Leakey, R.E.F. (1976). “New hominid fossils from the Koobi Fora formation in Northern Kenya.” Nature. 261:574-576. <>

The Open Source Paleontologist (2008). “3D Slicer: The Tutorial Parts I-VI.” <>

Ratiu, P., et al. (2004). “The tale of Phineas Gage, digitally remastered.” Journal of Neurotrauma. 21(5):637-643. <>

Russo, G.A. and E.C. Kirk (2013). "Foramen magnum position in bipedal mammals." Journal of Human Evolution. 65:656-670. <>

Russo, G.A. and E.C. Kirk (2017). "Another look at the foramen magnum in bipedal mammals." Journal of Human Evolution. 105:24-40. <>

Ruth, A.A., et al. (2016). “Locomotor pattern fails to predict foramen magnum angle in rodents, strepsirrhine primates and marsupials.” Journal of Human Evolution. 94:45-52. <>

Sartono, S. (1972). “Discovery of Another Hominid Skull at Sangiran, Central Java.” Current Anthropology. 13(1):125-126. <>

Schindelin, J., et al. (2012). "Fiji: an open-source platform for biological-image analysis." Nature Methods. 9(7):676-682. <>

Smithsonian National Museum of Natural History (2018). “Human Fossil Collection: Fossil Specimens.” <>

Timothy, F.C.D. (2006). "The Foramen Magnum: How Do We Know?" Afarensis: Anthropology, Evolution, and Science Blog. <>

Woodward, A.S. (1921). “A New Cave Man from Rhodesia, South Africa.” Nature. 108:371-372. <>

Data “African Fossils Virtual Laboratory.” <>

Bauer, Eric, Biology Professor Elon University, NC. “Human Fetal Skull.” <>

Boston Children’s Hospital Laboratory of Cognitive Neuroscience. “Skull of Phineas Gage.” <>

Dilmen, Nevit, Radiologist Sonomed Tibbi Goruntuleme Merkezi, Istanbul, Turkey. “Skull with overbite from CT data.” <>

Duke University Evolutionary Anthropology. “MorphoSource.” <>

Embodi3D. “Medical 3D Printing.” <>

Harvard University Museum of Comparative Zoology. “MCZBase: The Database of the Zoological Collections.” <>

Human Evolution Teaching Materials Project. “3D Models.” <>

Idaho State University Idaho Museum of Natural History, “Idaho Virtualization Laboratory.” <>

Kyoto University Primate Research Institute (KUPRI). “Digital Morphology Museum.” <>

Ohio University Witmer Lab. “Human skull, air sinuses, inner ear, and brain endocast.” <>

Radiolab. “Taung Child Skull.” <>


3D Slicer: A multi-platform, free and open source software package for visualization and medical image computing. <>

Adobe Acrobat Reader DC. <>

Blender: Open Source 3D Creation. <>

Fiji: Fiji Is Just ImageJ. <>

GIMP: GNU Image Manipulation Program. <>

Gnumeric: An open-source spreadsheet program. <>

Gpaint: GNU Paint. <>

ImageJ: An open source platform for scientific image analysis. <>

LibreOffice Calc: Spreadsheet. <>

MeshLab: The open source system for processing and editing 3D triangular meshes. <>


I would like to thank my father for teaching me anthropology and giving me advice and encouragement throughout my work. I thank my mother for finding skull models sources and references. I also thank her for her advice and persistent encouragement; I would not have done this project without her. I thank them both for their help editing my work.

I would like to thank the following institutions and individuals for making their specimens freely available on the Internet: Harvard University Museum of Comparative Zoology,, Duke University Evolutionary Anthropology, Kyoto University Primate Research Institute, Idaho Museum of Natural History, Witmer Lab at Ohio University, Boston Children’s Hospital Laboratory of Cognitive Neuroscience, Radiolab, Eric Bauer of Elon University, and Nevit Dilmen of Sonomed Tibbi Goruntuleme Merkezi.

I would like to thank the members of the open source software community, especially the developers of the following: Debian Linux, MX Linux, 3D Slicer, GIMP, ImageJ, Gnumeric, LibreOffice, MeshLab, Blender, and Gpaint.