UC Berkeley

Abstract

Cranial Morphology, Variation, and Integration in Homo sapiens

by

Whitney Brooke Reiner,

Doctor of Philosophy in Integrative Biology

University of California, Berkeley

Professor Leslea Hlusko, Chair

Herein I present three separate manuscripts pertaining to cranial morphology, variation, and integration in humans. The first manuscript introduces a newly recovered partial calvaria, OH 83, from the upper Ndutu Beds of Olduvai Gorge, Tanzania. I present the geological context of its discovery, a comparative analysis of its morphology, and place OH 83 within the context of our current understanding of the origins and evolution of Homo sapiens. The morphology of OH 83 was analyzed using quantitative and qualitative data from penecontemporaneous fossils and the W.W. Howells modern human craniometric dataset.

OH 83 is geologically dated to ca. 60–32 ka. Its morphology is indicative of an early modern human, falling at the low end of the range of variation for post-orbital cranial breadth, the high end of the range for bifrontal breadth, and near average in frontal length.

There have been numerous attempts to use cranial anatomy to define the species Homo sapiens and identify it in the fossil record. These efforts have not met wide agreement by the scientific community due, in part, to the mosaic patterns of cranial variation represented by the fossils. The variable, mosaic pattern of trait expression in the crania of Middle and Late Pleistocene fossils implies that morphological modernity did not occur at once. However, OH 83 demonstrates that by ca. 60–32 ka modern humans in Africa included individuals that are at the fairly small and gracile range of modern human variation.

In the second manuscript I provide craniometric data from Early Period (ca. 5000 B.P.) hunter-gatherers from the Sacramento Valley and the San Francisco Bay Area that represent some of the earliest indigenous Californians. I compare these data to the published worldwide human craniometric data set to provide perspectives on the range of human variation and the inter-relatedness of that variation.

I collected 76 cranial measurements and five indices from 59 adult crania collected using a three-dimensional (3D) digitizer (MicroScribe G2, Immersion Corporation), following published protocols associated with the comparative data set.

I conducted two sets of analyses exploring the range of variation, and calculating correlations. My analyses reveal that the Early Period Native Californians extends the known range of variation for 20 measurements. For six of the measurements, the smaller end of the range is extended, while the higher end of the range is extended for 14 measurements. For Native Americans, the Early Period Native Californians extend the range for 53 measurements, four of which are extended at both ends of the range. Correlation matrices for these data suggest the face is an integrated region of the cranium across modern humans, but specific patterns of correlation within and between regions of the cranium varied across populations. The early Native Californian crania exhibited the strongest overall correlations, differing significantly from the other samples (Mantel test, p < .0001).

Bringing the Early Period Native California morphologies into the published W.W. Howells data set provides an improved appreciation of the range of cranial variation in modern humans. While the message of Howells’ assessment that modern human crania vary widely was well-established by Howells, the Early Period data underscore this point. The evidence for integration within the facial skeleton revealed by the correlation matrices observed across all populations corroborates previous research demonstrating that the mammalian facial skeleton is an integrated region that varies fairly independently from the rest of the cranium.

In the third and final manuscript, I explore the influence of sample composition on the patterns of correlation for modern human crania by assessing correlation patterns at the level of the species, geographic regions, and populations, and the variation in sexual dimorphism at each of these levels.

I analyzed patterns of correlation for craniometric traits using the W.W. Howells’ worldwide human craniometric data set, and data I collected from 59 adult Early Period Native Californian crania (ca. 5000 B.P.) using a three-dimensional (3D) digitizer

(MicroScribe G2, Immersion Corporation) following Howells’ definitions to locate all anatomical landmarks and collect all measurements. Using these data, I generated correlation matrices for samples of varying composition to test three hypotheses for cranial integration. I test these hypotheses at the level of the species, then divided by geographic region, and at the population level, and further decompose each of these samples by sex.

I found that patterns of correlation varied the most at the population level. Patterns of correlation did not persist at all levels below the species. Thus, sample composition does influence patterns of correlation, but is inversely proportional to the sample’s level of composition, such that variation between samples is greater between populations than it is between geographic regions.

Studies investigating human cranial integration are often interpreted at the species level, even when based on analyses of below-species level data. Because the influence of sample composition on these analyses is not well-known, it is unclear if this broad application of the data is merited. My results demonstrate that patterns of correlation may reveal signals of the underlying mechanisms responsible for generating sexual dimorphism and inter-population variation in cranial morphology as a dataset becomes more specific. These patterns of variation may stem from environmental pressures that could influence cranial shape, and even development via epigenetic interactions. While the use of pooled-sex samples of multiple populations could reduce population-specific noise, by design, it could also temper signals of biologically interesting and informative patterns within and between populations. Furthermore, understanding the way patterns of correlation vary at the level of the population could benefit researchers seeking to combine data from multiple populations into a single sample with the least amount of population-specific environmental pressures or epigenetic interactions that could bias the data. This study represents an initial effort to begin to understand this variation.