Relationship of cranial robusticity to cranial form, geography and climate

UPDATE (June 30)
This paper investigates the relationship between cranial robusticity and a number of factors said to underlie it, including cranial size and shape, climate, and neutral genetic variation. Genetic similarity between populations was assessed using the well-known Rosenberg et al. dataset from 2002.

From the paper:

If the robusticity traits are the subject of neutral evolutionary processes, then the distance matrix based on these characters will be strongly correlated with that based on the neutral genetic markers (microsatellite data) (e.g., Roseman, 2004).

...

A functional hypothesis that specifically implicates forces associated with mastication would be supported by a stronger correlation between cranial robusticity and the MLS rather than CLS as the former more directly captures morphology associated with mastication, although it is also possible that changes in overall
cranial shape may be related to mastication. A strong relationship between cranial robusticity and the climatic variables would support the influence of the local environment on the development of cranial robusticity.

Note: CLS/MLS=cranial/masticatory landmark set.

First of all, it turns out that Robusticity traits are closely correlated with each other, suggesting that they do indeed capture an overall factor of "Robusticity" rather than being independent from each other. A notable exception is occipital torus, which is not significantly related to other robusticity traits.

Next up, principal components analysis was performed:

In deciding how many PCs to evaluate, we applied the common Guttman-Kaiser criterion (keep all PCs with eigenvalues [1.0; Kaiser, 1961), which results in the retention of the first three components. However, a more conservative criterion, the Scree Plot (Cattell, 1966), suggests that only PC 1 should be retained. Although PC 1 accounted for a proportionally larger percentage of the total variance in cranial robusticity (27%), the second and third components each explain 11% of the variance and may indicate that there is more than one relevantpattern of cranial robusticity (Table 10).

...

The first PC reflects overall levels of robusticity as all 11 traits load positively (Table 10), although the occipital torus has a loading near zero. ... The groups with the highest median (and mean) scores are New Zealand, Australia/Tasmania, North
America, and South America, while the lowest scores belong to Mongolia, East Asia, Inuit, and Khoe-San (see Fig. 3).

...

Males score significantly higher than females on PC 1 within all groups except the East and West Europeans, East and West Africans, and Khoe-San (Table 11), but males scored higher on average than females even in those groups that did not reach statistical significance

Interestingly, in the test for sex differences in PC1, Europeans differed from each other. Southern Europeans (Peloponnesian Greeks and Italians) were most dimorphic, and West Europeans (Austrians and Germans) were least.

From the paper:

The second PC has both high positive (occipital torus, rounding of orbits) and high negative (sagittal keel, anterior mastoid, bregmatic eminence) loadings ... The highest scoring groups on PC 2 are East Asia, Mongolia, Australia/Tasmania, and the Khoe-San, while East Africa, West Africa, and Eastern Europe have low scores (Fig. 3a). Many of the pair-wise contrasts between the highest and lowest scoring groups are significant, particularly those that include East Africa, East Asia, and Mongolia (Table 10). It appears that while the East Asian populations are gracile overall (see above), they do display some characters typically considered as ‘‘robust’’ (e.g., the occipital torus).

...

The third PC also has a mixed pattern of positive and negative loadings. The traits with the highest loadings are sagittal keel, occipital torus, malar tubercle, bregmatic
eminence (all positive), infraglabellar notch and supraorbital torus (both negative). The North American and, to a lesser extent, New Zealand, groups score highest
on PC 3, in contrast to Australia/Tasmania, Southern Europe, Eastern Europe, and East Africa (Fig. 3b).

In PC1, sex and size differences contribute about 36% of the variation, but only 3 and 8% in PC2 an PC3 respectively.

The author calculated distances between populations for Robusticity, CLS, MLS, Climate, and microsatellites, and sees how the inter-population distance based on robusticity correlates with the other four potentially explanatory factors:

The correlation coefficients from the Mantel tests are weak, ranging from -0.115 to 0.387 (Table 12). The null hypothesis of neutral evolution was rejected as the robusticity distances were not significantly correlated with neutral genetic distances.

The strongest (and only significant) correlations are between cranial robusticity and cranial (CLS) or masticatory apparatus shape (MLS). Cranial robusticity in the combined male–female sample is significantly correlated with the masticatory shape, and its correlation with overall cranial shape approached significance (Table 12).

On the Southern European masticatory system:

Whereas South Europe is among the lowest scoring groups on both the shape and robusticity vectors, the highest scoring groups on the robusticity vector (e.g., South America and New Zealand) are not the highest scoring groups on the shape vector (specifically Australia/Tasmania). The more gracile groups (e.g., South European) have more anteriorly positioned zygomatic bones (as indicated by the inferior zygomaticotemporal suture and zygomaxillare), more laterally located postglenoid processes and frontotemporale, and relatively larger cheek teeth (in the anteroposterior direction) that are more superiorly positioned.

Finally, a bit on Australian aboriginals who often get singled out as being particularly robust. It turns out that they are, but their pattern of robusticity involves particular traits, while other human groups, such as Native Americans are robust in a different manner:

While Aboriginal Australians have long been the standard bearers for robust cranial morphology, this study reveals that human populations exhibit more than one pattern of cranial robusticity. The results of this study emphasize a primary trend of variability from gracile to robust (except in the occipital torus region), but also highlight secondary patterns of differential cranial trait expression within populations. For example, the Native American group from Grand Gulch, Utah is characterized by robust expression of the sagittal keel, bregmatic eminence, occipital torus, and malar tubercle, but a more gracile supraorbital region in contrast to the pattern seen in Aboriginal Australians.

All in all, this is an excellent data-driven paper, which combines data from skulls, genes, and climate to arrive at a comprehensive study of the phenomenon of modern human cranial robusticity and its etiology.


American Journal of Physical Anthropology doi:10.1002/ajpa.21120

Relationship of cranial robusticity to cranial form, geography and climate in Homo sapiens

Karen L. Baab et al.

Abstract

Variation in cranial robusticity among modern human populations is widely acknowledged but not well-understood. While the use of robust cranial traits in hominin systematics and phylogeny suggests that these characters are strongly heritable, this hypothesis has not been tested. Alternatively, cranial robusticity may be a response to differences in diet/mastication or it may be an adaptation to cold, harsh environments. This study quantifies the distribution of cranial robusticity in 14 geographically widespread human populations, and correlates this variation with climatic variables, neutral genetic distances, cranial size, and cranial shape. With the exception of the occipital torus region, all traits were positively correlated with each other, suggesting that they should not be treated as individual characters. While males are more robust than females within each of the populations, among the independent variables (cranial shape, size, climate, and neutral genetic distances), only shape is significantly correlated with inter-population differences in robusticity. Two-block partial least-squares analysis was used to explore the relationship between cranial shape (captured by three-dimensional landmark data) and robusticity across individuals. Weak support was found for the hypothesis that robusticity was related to mastication as the shape associated with greater robusticity was similar to that described for groups that ate harder-to-process diets. Specifically, crania with more prognathic faces, expanded glabellar and occipital regions, and (slightly) longer skulls were more robust than those with rounder vaults and more orthognathic faces. However, groups with more mechanically demanding diets (hunter-gatherers) were not always more robust than groups practicing some form of agriculture.

Link