Measuring cranial variation using geography as a proxy for neutral genetic distances

world globe flushing
Certain anatomical features of the human skeleton are known to vary with geography and climate. To what extent each variable contributes to our physical makeup is less clear. The problem is that populations with similar climate are geographically close to one another. Even if we find shared traits among populations from similar climates it may be just as a result of geographic proximity (and thus clinal gene flow), rather than shared common ancestry.

As I mentioned in my previous post, anthropologists often compared cranial data to matched microsatellite datasets. However, it is rarely possible to get an exact match between the cranial and microsatellite populations. The anthropologist will instead use populations that are genetically similar and which may or may not be representative of the target population. Another option is to substitute microsatellite data with geographic distances, since studies have found a strong correlation between genetic distance and geographic distance (Manica et al. 2005; Ramachandran et al. 2005; Romero et al. 2008). This allows us to get around the need to match phenotypic data with genetic datasets.

A recent paper by Betti et al. used geographic distance as a proxy for neutral genetic distance. They set out to test the extent to which cranial differences can be explained by geographic proximity, by comparing pairwise phenotypic distances among populations and pairwise geographic distances using isolation by distance (IBD) models, as well as comparing pairwise cranial distances with climatic variables after correcting for IBD. Geographic distances were calculated as the shortest distance over land between populations while avoiding areas greater than 2000 metres above sea level. Intercontinental land bridges were also factored into their model.

Their study found geographic distance (and by extension genetic distance) to be a strong predictor of cranial variation. Minimum and maximum temperatures were also a significant predictor of cranial differentiation but not as strong as geographic distance. It also appears that much of this climate-related variation is influenced by the populations from exceptionally cold climates. A previous study by Roseman also found that populations living in extremely cold climates showed greater selection. Betti et al. suggest that this may be due to culture acting as an environmental buffer, with the buffer breaking down at extremely cold climates, after which cranial plasticity takes over.

Since climate and geographic distance covary, not considering isolation by distance leads to an overestimation of the effect of climate on cranial differences between populations. Not surprisingly facial traits showed the strongest correlation with climate. In summary, this study suggests that cranial measurements are predominately influenced by neutral evolutionary processes, especially in populations that do not live in extremely cold climates.

Betti et al. 2009. The relative role of drift and selection in shaping the human skull. Am. J. Phys. Anthropol. in press.

Manica A, Prugnolle F, Balloux F. 2005. Geography is a better determinant of human genetic differentiation than ethnicity. Hum Genet 118:366–371.

Ramachandran S, Deshpande O, Roseman CC, Rosenberg NA, Feldman MW, Cavalli-Sforza LL. 2005. Support from the relationship of genetic and geographic distance in human populations for a serial founder effect originating in Africa. Proc Natl Acad Sci USA 102:15942–15947.

Romero IG, Manica A, Goudet J, Handley LL, Balloux F. 2008. How accurate is the current picture of human genetic variation? Heredity 102:120–126.

Bookmark and Share
Above photo by jacorbett70 under creative commons license.
View Comments

Examining cranial robusticity

robust hominin
Palaeoanthropologist Darren Curnoe (2009) gives the following biological definition of the term ‘robust’:

…a descriptive anatomical term referring to individuals, complexes, organs, structures or traits which are heavily built, rugged, well defined or corpulent.

Bones tend to more robust where muscles, tendons or ligaments insert into the periosteum. When these insertion sites are subjected to stress, blood flow increases. This in turn stimulates the production of osteoblasts, which lay down extra bone. With respect to the skull the term robust is generally used to refer to so-called superstructures, such as the supraorbital ridges, occipital crests or zygomaxillary tuberosities. Anthropologists often classify robusticity based on the relative expression of a particular trait, or indeed its absence. Given that robusticity is related to physical stress, traits tend to be more pronounced in males and in certain populations (e.g. Aboriginal Australians and Fuegians).

The retention of robust features in certain populations, particularly Aboriginal Australians, has been used to support the multiregional hypothesis of human origins (e.g. Wolpoff et al. 2001; Frayer et al., 1993). On the other hand, proponents of a replacement model see robust traits (e.g. in Australian Aboriginal populations) as retained plesiomorphies and argue that these traits cannot be used to show continuity (Lieberman 2000). In response, many multiregionalists have revised their position to suggest that the reduction of the browridge in later Neandertals, such as St Césaire and Vindija, represents a synapomorphy between Neandertals and modern humans, likely due to interbreeding. The underlying assumption here is that these robust traits have a strong genetic component. Furthermore, there is a notable decrease in cranial robusticity from the early Upper Palaeolithic to late Upper Palaeolithic. It has been suggested that this may reflect changes in diet. Transition from hunter-gather to agricultural lifestyle is associated with a reduction in cranial robusticity, although correlation does not necessarily prove causation. However, not all hunter-gather groups are universally more robust than argriculturalists, which might suggest some other factors at play.

A recent in press paper by Baab et al. sets out to examine the possible mechanisms behind robust cranial characters. The null hypothesis in their study is that neutral evolutionary processes (e.g. genetic drift) were responsible for the pattern of cranial robusticity in modern humans - the rejection of which would suggest selection acting on these traits. To test the null hypothesis of neutral evolution of cranial robusticity Mahalonobis D2 distances for robust characters were compared to Ddm distances derived from microsatellite data. Microsatellites are useful in reconstructing evolutionary relationships due to their unusually high mutation rates, which result in largely selectively neutral polymorphisms.

Of the variables examined, only cranial shape was significantly correlated with robusticity, while cranial size, climate and neutral genetic distances were not. This is at odds with an earlier study by Mirazón Lahr and Wright (1996) (1996) who found the strongest correlation between cranial robusticity and cranial size. This finding may be due to use of geometric morphometrics by Baab and colleagues, which is better at separating size and shape compared to the linear morphometrics used by Mirazón Lahr and Wright (1996).

Cranial robusticity was not correlated with neutral genetic distances, suggesting that neutral evolutionary processes (e.g. genetic drift) were not responsible for the pattern of cranial robusticity in the populations studied. As noted by the authors, this finding could also be explained by a non-perfect match of populations among some of the cranial and molecular samples. In studies such as this one, it is often difficult to find an exact match between the populations from which we derive our cranial and molecular data. In such cases, we are left with the choice of eliminating samples or using another genetically similar population. The authors choose the latter but neither option is ideal and both have their own disadvantages. Unfortunately, the reason for including Upper Palaeolithic and Neolithic samples in this study is never fully explained and the assumption that modern genetic populations are appropriate proxies for such populations is never justified. Setting this aside, the findings of this study caution the use of robust traits in constructing phylogenetic relationships in modern humans.

The strongest correlations were found between cranial robusticity and either cranial or masticatory shape. This lends support to the hypothesis that robusticity is in some part functionally determined. The study also found crania with more prognathic faces, longer skulls, expanded glabellar and occipital regions to be more robust. Mirazón Lahr and Wright (1996) noted a similar tendency of longer skulls to have superstructures, while further emphasising their tendency to be associated with narrow skulls and a large palatal region.

While most of the robust variables in this study were areas of muscle insertions, the supraorbital region has a distinct aetiology. While many have interpreted the supraorbital region as an area of stress reinforcement in the skull (the so-called beam model) which is strongly influenced by mastication (Endo 1966, 1970; Russell 1985), there is a strong evidence to suggest that this is not its primary purpose. Supraorbital development begins early in life, suggesting that the supraorbital ridge may be part of the overall craniofacial complex and is likely under genetic control. While the beam model is intuitive, it is unsupported by empirical data. Hylander and colleagues (Hylander et al. 1991a, 1991b, 1992; Hylander and Ravosa 1992) conducted in vivo strain gauge experiments in different primates to assess the amount of strain magnitudes generated during mastication. They found these levels to be low to induce bone deposition in all the species they studied, even when chewing hard food. Moreover, anthropoids do not show a correlation between the browridge and the moment arms of the masticatory muscles, as the beam model would predict (Ravosa 1991). These researchers adopt the model proposed by Moss and Young (1960), which views supraorbital development as the result of placement of the brain and eyes. They postulated that the reduction of the brow ridge in modern humans was related to the expansion of the frontal lobe in our species. In hominins with orbits positioned well in front of the frontal lobes, as in chimpanzees or the erectines, the space between the orbits and the brain case is bridged by a brow ridge. If the supraorbital region is under genetic control, as the research of Hylander and Ravosa suggests, it would be of interest to examine this region in isolation to assess if it correlates with neutral evolutionary processes, particularly in light of a recent paper by Von Cramen-Taubedal which found the shape of the frontal bone to be consistent with neutral genetic expectation.


Baab KL, SE Freidline, SL Wang, T Hanson. 2009. Relationship of cranial robusticity to cranial form, geography and climate in Homo sapiens (in press). Am. J. Phys. Anthropol.

Curnoe D. 2009. Possible causes and significance of cranial robusticity among Pleistocene-Early Holocene Australians. Journal of Archaeological Science (2009) vol. 36 (4): 980-990.

Endo B. 1966. Experimental studies on the mechanical significance of the form of the human facial skeleton. J Faculty Sci Univ Tokyo (Section V, Anthropol) 3:1–106.

Endo B. 1970. Analysis of stress around the orbit due to masseter and temporalis muscles respectively. J Anthropol Soc Nippon 78:251–266.

Frayer DW, MH Wolpoff, AG Thorne, FH Smith, GG Pope. Theories of modern human origins: the paleontological test. American Anthropologist (1993) vol. 95 (1): 14-50.

Hylander WL, Picq PG, Johnson KR. 1991a. Masticatory–stress hypotheses and the supraorbital region of primates. Am J Phys Anthropol 86:1–36.

Hylander WL, Picq PG, Johnson KR. 1991b. Function of the supraorbital region of primates. Arch Oral Biol 36:273– 281.

Hylander WL, Ravosa MJ. 1992. An analysis of the supraorbital region of primates: a morphometric and experimental approach. In: Smith P, Tchernov E,
editors. Structure, function and evolution of teeth. Tel Aviv: Freund Publishing. p 223–255.

Lieberman, DE. (2000) Ontogeny, homology, and phylogeny in the Hominid craniofacial skeleton: the problem of the browridge. In P. O'Higgins and M. Cohn (eds.) Development, Growth and Evolution: implications for the study of hominid skeletal evolution. London: Academic Press, pp. 85-122.

Moss ML, RW Young. 1960. A functional approach to craniology. Am. J. Phys. Anthropol. 18:281-292.

Mirazón Lahr M, RVS Wright. 1996. The question of robusticity and the relationship between cranial size and shape in Homo sapiens. Journal of Human Evolution.

Ravosa MJ. 1991. Interspecific perspective on mechanical and nonmechanical models of primate circumorbital morphology. Am J Phys Anthropol. 86(3):369-96.

Russell MD. 1985. The Supraorbital Torus:" A Most Remarkable Peculiarity". Current Anthropology. vol. 26 (3) pp. 337

Wolpoff MH, J Hawks, DW Frayer, K Hunley. 2001. Modern Human Ancestry at the Peripheries: A Test of the Replacement Theory. Science. vol. 291 (5502):293-297.

Bookmark and Share
Above photo modified from original by Thomas Hawk under creative commons license.
View Comments

The impact of cranial plasticity on the reconstruction of human population history

Palaeoanthropologists are interested in reconstructing the evolutionary history of our species from fossil remains. Like other taxonomists they are primarily interested in identifying shared derived traits, otherwise known as synapomorphies. A synapomorphic trait is shared by some members of a taxa and not by others, since the former inherited this trait from a common ancestor. For instance, the retention of a clavicle (collar bone) is one of the more cited synapomorphic traits common to all primates.

A homoplasy is a trait that is present in two or more taxa but that has not been derived through common ancestry but rather through convergence, parallelism, or reversal. The wings of insects, birds and bats are homoplasies, since they arose through convergent evolution. Thus, homoplasies and synapomorphies may be identical in appearance but are distinguished by whether or not they arose through common ancestry. As a result, it can often difficult to pry apart traits which are synapomorphies from those which are homoplasies. A subset of homoplasies are termed homoiologies. Lycett and Collard (2005) define homoiologies as:

“… phylogenetically misleading resemblances among a group of taxa that can be ascribed to phenotypic plasticity. That is, homoiologies are homoplasies that result from the expression by a genotype of different phenotypes in response to different environmental conditions.”

They arise primarily from nonheritable epigenetic responses to mechanical stimuli. The "homoiology hypothesis" (Lieberman 1995) was derived from the well known fact that bone shape and size can be modified by mechanical loading. As such, homoiologies are expected to have greater influence upon the more plastic regions of a phenotype.

The homoiology hypothesis makes two testable predictions:

1. Traits subject to biomechanical stress should exhibit higher within-taxon variability due to the increased plasticity.

2. These traits should be less reliable for reconstructing phylogeny.

Previous studies have examined the homoiology hypothesis in various primate species (Collard and Wood 2007, 2000). These studies found that the regions of the skull associated with mastication indeed exhibited higher within-taxon variability but they were as reliable in reconstructing phylogenetic relationships as other regions of cranium not directly associated with masticatory function. Since these studies looked at interspecific studies, it was suggested that maybe homoiology was a greater problem for intraspecific studies.

A paper (in press) in the Journal of Human Evolution by Noreen von Cramon-Taubadel tests the homoiology hypothesis in an intraspecific study of human populations. Areas of the skull related with mastication would be expected to be under greater biomechanical stress, and as such be more affected by homoiology. She divided the skull into zones thought to be related to mastication (zygotemporal and palatomaxilla regions) and zones relatively unaffected by mastication (the upper face, cranial vault and basicranium). She tested the predictions of the homoiology hypothesis by comparing craniometric data with matched molecular data for 12 modern human populations.

Like previous interspecific studies, regions of the skull related to mastication show great variability (as predicted by the homoiology hypothesis) but these regions were no less reliable at reconstructing phylogenies (at variance with the homoiology hypothesis). It is worth noting that if biomechanical stress affects all individuals in the same way then these characters will not confound the phylogenetic analysis.

These findings mean that the homoiology hypothesis is flawed in at least some of its premises. The results of this and previous studies suggest that within-taxon variability should not be used to assertion the usefulness of cranial traits for determining phylogenetic relationships. Moreover, it does not appear to hold that homoiologies are any more problematic in determining intraspecific evolutionary relationships as interspecific ones. Finally, even though regions of the skull related to mastication are more variable than non-masticatory regions, they do not seem to be any less reliable for the reconstruction of phylogenies.

Collard and Wood. Hominin homoiology: An assessment of the impact of phenotypic plasticity on phylogenetic analyses of humans and their fossil relatives. Journal of human evolution (2007) vol. 52 (5) pp. 573-584.

Collard and Wood. How reliable are human phylogenetic hypotheses?. Proc. Natl. Acad. Sci. U.S.A. (2000) vol. 97 (9) pp. 5003-6.

Cramon-Taubadel. Revisiting the homoiology hypothesis: the impact of phenotypic plasticity on the reconstruction of human population history from craniometric data. Journal of Human Evolution (2009) pp. 1-12.

Lieberman. Testing hypotheses about recent human evolution from skulls: integrating morphology, function, development, and phylogeny. Curr. Anthropol. (1995) 36, 159–197.

Lycett and Collard. Do homoiologies impede phylogenetic analyses of the fossil hominids? An assessment based on extant papionin craniodental morphology. Journal of human evolution (2005) vol. 49 (5) pp. 618-642.

Wood and Lieberman. Craniodental variation in Paranthropus boisei: a developmental and functional perspective. American Journal of Physical Anthropology (2001) vol. 116 (1) pp. 13-25.

Bookmark and Share
Above photo by wauter de tuinkabouter under creative commons license.
View Comments

Race: an anthropological perspective

Forensic anthropologists often have to establish a biological profile of an individual based on their skeletal remains. This involves establishing probable age, sex, stature and race. The last category seems to be at odds with the consensus view in modern biology, which views the race concept to be biologically meaningless. Since this is the case, then why do forensic anthropologists insist on determining race?

Alice Brues defined race as "a division of a species which differs from other divisions by the frequency with which certain hereditary traits appear among its members." This definition of race, like most others, is rather equivocal, in that it does not tell us how much variance in the frequency of traits necessitates the creation of a new race. If we take this definition at face value then according to craniometric and genetic data an incalculable number of races exist.

In 1972, R.C. Lewontin reported that, for genes at a single locus, most genetic variation existed within populations, rather than between them. For most biologists this put the nail in the coffin for the race concept. In a re-examination of Lewontin's findings, the Cambridge statistician A.W.F. Edwards, noted that our ability to correctly classify populations is due to the correlations among different loci. By focusing on multiple loci the between population differences increase dramatically.

As way of an analogy, imagine we asked a stranger the following question via internet: "Which colour do you prefer less: orange or brown?" Previous surveys have showed that there is a slightly greater tendency for women to rate orange as their least favourite colour, while men have a slightly greater tendency to dislike brown. However, there is a very high amount of overlap. We would have a tough time trying to predict sex based on the results of this single question. However, if we ask say thirty questions instead of one, we would be able to predict sex with a much higher degree of certainty based on the responses as a whole. In a similar manner, we could not confidently determine race of an individual based on one or two cranial measurements. However, the likelihood of a positive determination increases significantly when we include more measurements.

Craniometrics has been shown to correctly classify individuals into a few broadly defined racial categories, as well as many more geographically localised categories. The ability of forensic anthropologists to accurately classify individuals into predefined groups does not substantiate the biological race concept. Just because we can determine a skeleton to be of Irish, Western European, Northwestern European or European ancestry does not mean that such ancestral groups exist in any meaningful biological sense. However, such information is useful for homicide investigators who are interested in whittling down their list of possible missing persons.

So how are we to understand race? Human variation is probably best understood in terms of both temporal and geographic distances. Cranial variation correlates strongly with geography; meaning that the further apart the populations are geographically, the more dissimilar they are phenotypically. Conversely, neighbouring populations show greater phenotypic similarities, spurring anthropologist Frank Livingstone to write in 1962 "there are no races, only clines”. The relationship between phenotypic variation and geography is likely due to both isolation by distance (there is greater gene flow between neighbouring populations) and the many founder effects that occurred in the course of human history. The longer groups remain isolated the more dissimilar their genotypes. Since most racial categories are defined by geographic regions, it should not come as a surprise that there is a correlation between race and place of ancestry.

Race is a crude sociocultural construct based on the underlying reality of biological variation. In this regard it is similar to other cultural phenomena, which help us understand our past. For example, much can be ascertained about ancestry and human migration by studying languages. In this regard, race has proved to be a useful concept in the fields of medicine and law enforcement. As long as law enforcement continue to use racially defined categories, forensic anthropologists will similarly follow suit.

Bookmark and Share
Above photo modified from original by indianfilipino under creative commons license.
View Comments

Is craniometry scientific?

Craniometry is the measurement of the bones of the skull. The technique is primarily used to determine ancestry, population distances and sex. It differs greatly from the psuedoscientific fields of phrenology and physiognomy, although in the general populace the distinction is not so clear. The book Postmodernism and Race has the following definition: "Craniometry is the measurement of human skulls as an indication of intelligence", while a recent blog article on craniometry states that it "claims to be able to predict the intelligence and nature of human being." Craniometry says nothing about intelligence or human nature.

Craniometric studies were, to a large degree, racially motivated in the early decades of the twentieth century, with anthropologists trying to validate their preconceived racial categories. In 1912, Franz Boas published a study challenging the prevailing notion that certain cranial measurements were under ironclad genetic control. He studied the head form of some 13,000 European immigrants and their American-born children. He found significant differences in the shape of the heads between parents and their children, which he interpreted as evidence for cranial plasticity. In other words, environment, not genetics, shapes cranial morphology. During the subsequent decades, Boas' results came to be largely accepted by the anthropological community, with students of the Boasian school disregarding craniometric studies as an ill-fated enterprise. For many years, anthropologists steered clear of craniometry, instead focusing their expertise in other less stigmatised areas of physical anthropology such as palaeopathology.

Quite recently papers by Sparks
et al (2002) and Gravlee et al (2003) have re-examined Boas' original data. However, those expecting the final word on cranial plasticity were to be disappointed. The anthropologist Milford Wolpoff is quoted as saying in 1975 "The data do not speak for themselves. I have been in rooms with data and listened very carefully. They never said a word." This is particularly true of these two papers, which use the very same data to come to divergent conclusion. While Gravlee et al believe Boas to be essential correct, Sparks et al came down firmly on the other side. The real answer, I believe, is to be found between the lines.

Boas did, in fact, find a statistically significant environmental effect in his study but this begs the question of whether it is a meaningful effect. Sparks suggests that while the effect is real, it only constitutes a tiny proportion of variation. In fact, considering the size of Boas' sample (~13,000) it is almost impossible not to find statistically significant results; biology is, after all, intrinsically variable. It may be the case that Boas played up the importance of the environmental effects as a reaction to the racial thinking that was prevalent at that time.

Perhaps, the biggest problem with Boas' methodology was his reliance on only a handful of measurements and particularly the use of the cephalic index (ratio of head breadth to head length). Anders Retzius introduced the cephalic index as way of classifying skulls based on their overall shape. He defined three main categories: dolichocephalic (long headed), brachycephalic (broad headed) and mesocephalic (intermediate headed).

Most modern biological anthropologists are of the opinion that the use of a couple of measurements to describe a multi-complex structure such as the skull is absurd. Today, biological anthropologist will take dozens of measurements of the skull. W. W. Howells, who measured thousands of skulls from all over the world, had the following to say about the cephalc index: "When Anders Retzius, a century and half ago, invented the cranial index, he gave us an answer for which there was no question." Even Boas himself wrote the following in 1940: "Measurements should always have a biological significance. As soon as they lose their significance they lose also their descriptive value."

Craniometry is used today in biological anthropology as a means of determining the relationships of peoples through their phenotype. A phenotype is the visible manifestation of a genotype. Since there is rarely a one to one relationship between the genotype and phenotype we must first demonstrate that the phenotype is an accurate reflection of the genotype. If this is not the case craniometry would be no more scientific than phrenology. Narrow-sense heritability is the proportion of phenotypic variation that arises from only the additive genetic differences among individuals and is expressed as h
2 = VA/VP. Heribtaility is measured on a scale of 0 (no heritable variation) to 1 (all phenotypic variation is due to additive genetic effects). The average cranial h2 has been estimated at around 0.55 (Relethford 1994; Devor 1987). A heritability greater than 0.5 indicates that most phenotypic variation is the examined traits are attributable to genetic factors. Thus the proportionality of genotypic to phenotypic variance is a reasonable assumption. The true litmus test of any hypothesis is its predictive power. Craniometric data is used with surprising accuracy by forensic anthropologists to determine likely ancestry of unknown individual and by palaeoanthropologists to determine our relationship to other hominins. The modern scientific practice of craniometry distinguishes itself from psuedosciences like phrenology and physiognomy in that it is based on sound biological theory, it is testable, it is predictive and objective.

References cited
Boas F (1912) Changes in the bodily form of descendants of immigrants. American Anthropologist 14: 530-562.

Sparks CS, Jantz RL (2002) A reassessment of human cranial plasticity: Boas revisited. Proc. Natl. Acad. Sci. USA 99: 14636-14639.

Gravlee CC, Bernard HR, Leonard WR (2003) Heredity, Environment, and Cranial Form: A Reanalysis of Boas's Immigrant Data. American Anthropologist.

Devor EJ (1987) Transmission of human craniofacial dimensions. J Craniofac Genet Dev Biol 7: 95-106.

Relethford JH (1994) Craniometric variation among modern human populations. Am J Phys Anthropol 95: 53-62.

Bookmark and Share

View Comments
See Older Posts...