Stable carbon isotope data for early Pliocene hominins Ardipithecus ramidus and Australopithecus anamensis show narrow, C3-dominated isotopic signatures. Conversely, mid-Pliocene Au. afarensis has a wider isotopic distribution and consumed both C3 and C4 plants, indicating a transition to a broader dietary niche by ~ 3.5 million years ago (Ma). Dietary breadth is an important aspect of the modern human adaptive suite, but why hominins expanded their dietary niche ~ 3.5 Ma is poorly understood at present. Eastern Africa has produced a rich Pliocene record of hominin species and associated mammalian faunas that can be used to address this question. This dissertation hypothesizes that the shift in hominin dietary breadth was driven by a transition to more open and seasonal environments in which food resources were more patchily distributed both spatially and temporally. To this end, I use a multiproxy approach that combines hypsodonty, mesowear, faunal abundance, and stable isotope data for temporally well-constrained early and mid-Pliocene mammal assemblages (5.3-2.95 Ma) from Ethiopia, Kenya, and Tanzania to infer patterns of environmental change through time. Hypsodonty analyses revealed that early Pliocene sites had higher annual precipitation, lower precipitation seasonality, and lower temperature seasonality than mid-Pliocene sites. Mesowear analyses, however, did not show from attrition- to abrasion- dominated wear through time. Abundance data suggest that there was a trend towards aridity, as Tragelaphini (woodland antelope) decline while Alcelaphini (grassland antelope) increased in abundance through time. Carbon isotope data indicate that most taxa shifted to diets focusing on C4 grasses through time, which closely follows paleosol carbon isotope data documenting the expansion of grassland ecosystems in eastern Africa. Overall, the results suggest Ar. ramidus and Au. anamensis preferentially exploited habitats in which preferred food resources were likely available year-round, whereas Au. afarensis lived in more variable, seasonal environments in which preferred foods were available seasonally. Au. afarensis and K. platyops likely expanded their dietary niche in less stable environments, as reflected in their wider isotopic niche breadth.

Providing an environmental context to early hominins is as important as describing the hominin fossils themselves, because evolutionary processes are tightly linked to everchanging ecosystems that vary across space and through time. An optimal understanding of ecosystems changes is critical to formulate and test hypotheses regarding human evolution and adaptation. Fortunately, the fossil record has yielded abundant remains of mammals which can be used to explore the possible causal relationships between environmental change and mammal – including hominin –evolution. Although many studies have already been conducted on this topic, most of them are framed at large spatial and temporal scales. Instead, this dissertation focuses on the evolution and paleoecology of only one group of mammals (the Suidae) in a specific geographical area (lower Awash Valley in Ethiopia) and within a constrained time frame (3.8–2.6 Ma). Three dissertation papers address: 1) changes in suid taxonomic composition in relation to Late Pliocene faunal turnover ~2.8 Ma in the Lee-Adoyta basin, Ledi-Geraru; 2) comparisons of suid diets from Hadar (~3.45–2.95 Ma) with respect to those of Kanapoi (~4.1 Ma, West Turkana, Kenya); 3) the dietary ecology of the suids from Woranso-Mille (~3.8–3.2 Ma). Results of these papers show that 1) after ~2.8 Ma there is a replacement of suid species that is coupled with low relative abundance of suids. This is compatible with more open and/or arid environments at this time; 2) suid dietary breadth was broader in Hadar than in Kanapoi, but this is mostly driven by the dietary niche space occupied by Kolpochoerus in Hadar, a suid genus absent from Kanapoi; 3) suid diets vary both temporally and geographically within the lower Awash Valley. Kolpochoerus incorporates more C4 resources (e.g., grasses) in its diet after ~3.5 Ma and in general, suids after ~3.5 Ma in Woranso-Mille had C4-enriched diets in comparison with those from nearby Hadar and Dikika. Presumably, the changes in suid communities (relative abundance and taxonomic composition) and dietary shifts observed in suids were triggered by climatic and habitat changes that also contributed to shape the behavioural and morphological evolution of early hominins.

Across primates, molar-emergence age is strongly correlated to life-history variables, such as age-at-first-reproduction and longevity. This relationship allows for the reconstruction of life-history parameters in fossil primates. The mechanism responsible for modulating molar-emergence age is unknown, however. This dissertation uses a biomechanical model that accurately predicts the position of molars in adults to determine whether molar emergence is constrained by chewing biomechanics throughout ontogeny. A key aspect of chewing system configuration in adults is the position of molars: the distal-most molar is constrained to avoid tensile forces at the temporomandibular joint (TMJ). Using three-dimensional data from growth samples of 1258 skulls, representing 21 primate species, this research tested the hypothesis that the location and timing of molar emergence is constrained to avoid high and potentially dangerous tensile forces at the TMJ throughout growth. Results indicate that molars emerge in a predictable position to safeguard the TMJ during chewing. Factors related to the size of the buffer zone, a safety feature that creates greater stability at the TMJ during biting, account for a large portion of both ontogenetic and interspecific variation in the position of emergence. Furthermore, the rate at which space is made available in the jaws and the duration of jaw growth both determine the timing of molar emergence. Overall, this dissertation provides a mechanical and developmental model for explaining temporal and spatial variation in molar emergence and a framework for understanding how variation in the timing of molar emergence has evolved among primates. The findings suggest that life history is related to ages at molar emergence through its influence on the rate and duration of jaw growth. This dissertation provides support for the functionally integrated nature of craniofacial growth and has implications for the study of primate life history evolution and masticatory morphology in the fossil record.

Facial projection--i.e., the position of the upper face relative to the anterior cranial fossa--is an important component of craniofacial architecture in primates. Study of its variation is therefore important to understanding the bases of primate craniofacial form. Such research is relevant to studies of human evolution because the condition in

Homo sapiens--in which facial projection is highly reduced, with the facial skeleton located primarily inferior (rather than anterior) to the braincase--is derived vis-à-vis other primates species, including others in the genus Homo. Previous research suggested that variation in facial projection is explained by: (1) cranial base angulation; (2) upper

facial length; (3) anterior cranial base length; (4) anterior sphenoid length; and/or (5) anterior middle cranial fossa length. However, previous research was based on taxonomically narrow samples and relatively small sample sizes, and comparative data on facial projection in anthropoid primates, with which these observations could be

contextualized, do not currently exist.

This dissertation fills this gap in knowledge. Specifically, data corresponding to the hypotheses listed above were collected from radiographs from a sample of anthropoid primates (N = 37 species; 756 specimens) . These data were subjected to phylogenetically-controlled multiple regression analyses. In addition, multivariate and univariate models were statistically compared, and the position of Homo sapiens relative to univariate and multivariate regression models was evaluated.

The results suggest that upper facial length, anterior cranial base length, and, to a lesser extent, cranial base angle are the most important predictors of facial projection. Homo sapiens conforms to the patterns found in anthropoid primates, suggesting that these same factors explain the condition in this species. However, a consideration of the

evidence from the fossil record in the context of these findings suggests that upper facial length is the most likely cause of the extremely low degree of facial projection in Homo sapiens. These results downplay the role of the brain in shaping the form of the human cranium. Instead, these results suggest that reduction in facial skeleton size--which may

be due to changes in diet--may be more important than previously suggested.

Despite the critical role that the vertebral column plays in postural and locomotor behaviors, the functional morphology of the cervical region (i.e., the bony neck) remains poorly understood, particularly in comparison to that of the thoracic and lumbar sections. This dissertation tests the hypothesis that morphological variation in cervical vertebrae reflects differences in positional behavior (i.e., suspensory vs. nonsuspensory and orthograde vs. pronograde locomotion and postures). Specifically, this project addresses two broad research questions: (1) how does the morphology of cervical vertebrae vary with positional behavior and cranial morphology among primates and (2) where does fossil hominoid morphology fall within the context of the extant primates. Three biomechanical models were developed for the primate cervical spine and their predictions were tested by conducting a comparative analysis using a taxonomically and behaviorally diverse sample of primates. The results of these analyses were used to evaluate fossil hominoid morphology. The two biomechanical models relating vertebral shape to positional behaviors are not supported. However, a number of features distinguish behavioral groups. For example, the angle of the transverse process in relation to the cranial surface of the vertebral body--a trait hypothesized to reflect the deep spinal muscles' ability to extend and stabilize the neck--tends to be greater in pronograde species; this difference is in the opposite of the direction predicted by the biomechanical models. Other traits distinguish behavioral groups (e.g., spinous process length and cross-sectional area), but only in certain parts of the cervical column. The correlation of several vertebral features, especially transverse process length and pedicle cross-sectional area, with anterior cranial length supports the predictions made by the third model that links cervical morphology with head stabilization (i.e., head balancing). Fossil hominoid cervical remains indicate that the morphological pattern that characterizes modern humans was not present in Homo erectus or earlier hominins. These hominins are generally similar to apes in having larger neural arch cross-sectional areas and longer spinous processes than modern humans, likely indicating the presence of comparatively large nuchal muscles. The functional significance of this morphology remains unclear.

Extremely thick cranial vaults have been noted as a diagnostic characteristic of Homo erectus since the first fossil of the species was identified, but potential mechanisms underlying this seemingly unique trait have not been rigorously investigated. Cranial vault thickness (CVT) is not a monolithic trait, and the responsiveness of its layers to environmental stimuli is unknown. Identifying factors that affect CVT would be exceedingly valuable in teasing apart potential contributors to thick vaults in the Pleistocene. Four hypotheses were tested using CT scans of skulls of more than 1100 human and non-human primates. Data on total frontal, parietal, and occipital bone thickness and bone composition were collected to test the hypotheses: H1. CVT is an allometric consequence of brain or body size. H2. Thick cranial vaults are a response to long, low cranial vault shape. H3. High masticatory stress causes localized thickening of cranial vaults. H4. Activity-mediated systemic hormone levels affect CVT. Traditional comparative methods were used to identify features that covary with CVT across primates to establish behavior patterns that might correlate with thick cranial vaults. Secondly, novel experimental manipulation of a model organism, Mus musculus, was used to evaluate the relative plasticity of CVT. Finally, measures of CVT in fossil hominins were described and discussed in light of the extant comparative and experimental results. This dissertation reveals previously unknown variation among extant primates and humans and illustrates that Homo erectus is not entirely unique among primates in its CVT. The research suggests that it is very difficult to make a mouse grow a thick head, although it can be genetically programmed to have one. The project also identifies a possible hominin synapomorphy: high diploë ratios compared to non-human primates. It also found that extant humans differ from non-human primates in overall pattern of which cranial vault bones are thickest. What this project was unable to do was definitively provide an explanation for why and how Homo erectus grew thick skulls. Caution is required when using CVT as a diagnostic trait for Homo erectus, as the results presented here underscore the complexity inherent in its evolution and development.

The pattern and strength of genetic covariation is shaped by selection so that it is strong among functionally related characters and weak among functionally unrelated characters. Genetic covariation is expressed as phenotypic covariation within species and acts as a constraint on evolution by limiting the ability of linked characters to evolve independently of one another. Such linked characters are "constrained" and are expected to express covariation both within and among species. In this study, the pattern and magnitude of covariation among aspects of dental size and shape are investigated in anthropoid primates. Pleiotropy has been hypothesized to play a significant role in derivation of derived hominin morphologies. This study tests a series of hypotheses; including 1) that negative within- and among-species covariation exists between the anterior (incisors and canines) and postcanine teeth, 2) that covariation is strong and positive between the canines and incisors, 3) that there is a dimorphic pattern of within-species covariation and coevolution for characters of the canine honing complex, 4) that patterns of covariation are stable among anthropoids, and 5) that genetic constraints have been a strong bias on the diversification of anthropoid dental morphology. The study finds that patterns of variance-covariance are conserved among species. Despite these shared patterns of variance-covariance, dental diversification has frequently occurred along dimensions not aligned with the vector of genetic constraint. As regards the canine honing complex, there is no evidence for a difference in the pleiotropic organization or the coevolution of characters of the complex in males and females, which undermines arguments that the complex is selectively important only in males. Finally, there is no evidence for strong or negative pleiotropy between any dental characters, which falsifies hypotheses that predict such relationships between incisors and postcanine teeth or between the canines and the postcanine teeth.

Early hominins present an unusual pattern of sexual dimorphism. On one hand, the canine teeth of these species are weakly size-dimorphic, vertically short, and nonhoning, suggesting a social system characterized by infrequent, low-intensity intermale competition and monogamous pair-bonding. On the other hand, marked size variation in skeletal remains attributed to species of Australopithecus is thought to reflect strong body-mass dimorphism, which is more consistent with intense intermale competition. Reconciling these conflicting signals and understanding their adaptive significance is a major goal of paleoanthropology. This dissertation research contributes to this objective by investigating factors that may constrain or reduce canine height in extant anthropoid primates. Two hypotheses regarding the relationship between canine height and other elements of the masticatory system were tested using phylogenetic comparative methods. According to the first hypothesis, canine reduction is a pleiotropic by-product of changes in the sizes of other components of the dentition. With respect to canine height, the results of this study fail to support this idea. There is limited evidence for a relationship between basal canine crown dimensions and incisor and postcanine size, but significant interspecific correlations between these variables are not strong and are restricted primarily to the female maxillary dentition. These results indicate that if pleiotropy influences canine size, then its effects are weak. The second hypothesis proposes that canine reduction is a consequence of selection for increased jaw-muscle leverage. This hypothesis receives some support: there is a clear inverse relationship between canine height and the leverage of the masseter muscle in male anthropoids. Females do not exhibit this association due to the fact that dimorphism in muscle leverage is weak or absent in most anthropoid species; in other words, female muscle leverage tracks male muscle leverage, which is linked to canine height. Leverage of the temporalis muscle is not correlated with canine height in either sex. Two specimens of the 3.0-3.7-million-year-old hominin Australopithecus afarensis fall at or beyond the upper end of the great ape range of variation in masseter leverage, which is consistent with the idea that hominin canine evolution was influenced by selection for increased jaw-muscle leverage.