Citation
A comparative analysis of newly discovered Pliocene hominin footprints from Laetoli, Tanzania

Material Information

Title:
A comparative analysis of newly discovered Pliocene hominin footprints from Laetoli, Tanzania
Creator:
Pelissero, Alex J. ( author )
Place of Publication:
Denver, Colo.
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
1 electronic file (89 pages) : ;

Thesis/Dissertation Information

Degree:
Master's ( Master of science)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Anthropology, CU Denver
Degree Disciplines:
Anthropology

Subjects

Subjects / Keywords:
Bipedalism ( lcsh )
Human evolution ( lcsh )
Footprints, Fossil ( lcsh )
Bipedalism ( fast )
Footprints, Fossil ( fast )
Human evolution ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
The discovery of the 3.6 million year old Laetoli footprints was one of the most significant advances in our understanding of human evolution. In addition to confirming deep roots of bipedality as the defining characteristic of the hominin clade, its association with the contemporaneous species Australopithecus afarensis demonstrated that bipedality emerged well before encephalization in human evolution. Preserved trackways, like those seen at Laetoli, provide a crucial glimpse at both early hominin mobility and its paleoecological context. The recent discovery of additional, markedly larger, contemporaneous hominin footprints southwest of the Site G hominin footprints provides a critical comparative data set to test hypotheses about Pliocene hominin locomotion and ichnotaxonomy. Utilizing photogrammetric imagery and modeling, this thesis aims to compare the morphology of the two sets of trackways, as well as the gait patterns of the individuals responsible for creating the trackways, alongside modern human data. There are several affinities both morphologically and taphonomically between these two trackways, but the size differences seen between these two sets of trackways is likely associated with great differences in body size. This could be indicative of either significant size dimorphism and different locomotor repertoires in Au. afarensis, or potentially the presence of another hominin species at Laetoli. Situated within the broader paleoecological context of the Laetoli site, the qualitative and quantitative results of this analysis provides further insights into questions of early hominin taxonomic diversity and the evolution of upright bipedal locomotion.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: Adobe Reader.
Statement of Responsibility:
by Alex J. Pelissero.

Record Information

Source Institution:
University of Colorado Denver Collections
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
on10121 ( NOTIS )
1012124587 ( OCLC )
on1012124587
Classification:
LD1193.L43 2017m P46 ( lcc )

Downloads

This item has the following downloads:


Full Text
A COMPARATIVE ANALYSIS OF NEWLY DISCOVERED PLIOCENE HOMININ
FOOTPRINTS FROM LAETOLI, TANZANIA by
AlexJ. Pelissero
B.A., University of Colorado Boulder, 2010
A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Arts Anthropology Program
2017


2017
ALEX J. PELISSERO
ALL RIGHTS RESERVED


This thesis for the Master of Arts degree by
Alex Joel Pelissero has been approved for the Anthropology Program by
Charles Musiba, Chair Anna Warrener
Neffra Matthews


Pelissero, Alex J. (M. A., Anthropology)
A Comparative Analysis of Newly Discovered Pliocene Hominin Footprints from Laetoli Thesis directed by Associate Professor Charles M. Musiba
ABSTRACT
The discovery of the 3.6 million year old Laetoli footprints was one of the most significant advances in our understanding of human evolution. In addition to confirming deep roots of bipedality as the defining characteristic of the hominin clade, its association with the contemporaneous species Australopithecus afarensis demonstrated that bipedality emerged well before encephalization in human evolution. Preserved trackways, like those seen at Laetoli, provide a crucial glimpse at both early hominin mobility and its paleoecological context. The recent discovery of additional, markedly larger, contemporaneous hominin footprints southwest of the Site G hominin footprints provides a critical comparative data set to test hypotheses about Pliocene hominin locomotion and ichnotaxonomy. Utilizing photogrammetric imagery and modeling, this thesis aims to compare the morphology of the two sets of trackways, as well as the gait patterns of the individuals responsible for creating the trackways, alongside modern human data. There are several affinities both morphologically and taphonomically between these two trackways, but the size differences seen between these two sets of trackways is likely associated with great differences in body size. This could be indicative of either significant size dimorphism and different locomotor repertoires in Au. afarensis, or potentially the presence of another hominin species at Laetoli. Situated within the broader paleoecological context of the Laetoli site, the qualitative and quantitative results of this analysis provides further insights into questions of early hominin taxonomic diversity and the evolution of upright bipedal locomotion.
The form and content of this abstract are approved. I recommend its publication.
Charles Musiba
IV


TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION.........................................................1
II. BACKGROUND RESEARCH AND CURRENT LITERATURE..........................3
Theoretical Background............................................3
Current Literature................................................6
Geology and Paleoecology of Laetoli........................6
Locomotion of Laetoli Track-makers........................12
Locomotion of Australopithecus afarensis..................16
Pliocene Hominin Adaptive Radiation and Australopithecine Locomotor Variation.......................................22
III. MATERIALS AND METHODS.............................................29
Materials and Data Collection....................................29
Sample....................................................29
Descriptions of the Site S Hominin Footprints.............30
Photogrammetric Methods...................................37
Comparative Analysis......................................46
Metric Comparison and Analysis......................47
Procrustes Landmark Shape Analysis..................48
IV. RESULTS............................................................53
Comparison with other hominins and modern humans.................53
Kernel Density Plot and Procrustes Landmark PCA..................59
Summary of Results...............................................62
v


V. DISCUSSION...........................................63
VI. CONCLUSION..........................................71
REFERENCES...................................................73
VI


LIST OF TABLES
TABLE
3.1 Photogrammetric Capture Data.............................................39
3.2 Landmark Placement.......................................................49
4.1 Mean Metric Results......................................................53
vii


LIST OF FIGURES
FIGURE
2.1 Stratigraphy of Laetolil Beds....................................................7
2.2 Map of the Northern Tanzania Highlands..........................................9
2.3 Maps of Laetoli Region.........................................................10
2.4 Map of the Laetoli Site........................................................10
2.5 AL 333-15 partial foot.........................................................17
2.6 Burtele partial foot...........................................................21
3.1 TP2 Trench Footprints...........................................................33
3.2 M9 Trench Footprints...........................................................34
3.3 L8 Trench Footprints...........................................................35
3.4 L8 Trench Footprints...........................................................36
3.5 G Track Photogrammetric Model..................................................41
3.6 M9 Photogrammetric Model.......................................................42
3.7 TP2 Photogrammetric Model......................................................43
3.8 L8 Photogrammetric Model.......................................................44
3.9 Measurement Schemes............................................................47
3.10 S-1 Prints Used in 2-D Landmark Analysis......................................51
3.11 Site G Sample used in 2-D Landmark Analysis...................................52
4.1 Footprint Length Comparative Box-plots..........................................54
4.2 Step-length Comparative Box-plots..............................................55
4.3 Stature vs. Foot Length Plot...................................................56
4.4 Metric Data PCA Plot...........................................................57
4.5 DFA Plot.......................................................................58
viii
4.6 Landmark PCA Plot.
59


4.7 S-1 Print Kernel Density Plot.................................................60
4.8 G Avg. Kernel Density Plot....................................................60
4.9 2-D PCA Relative Deformations.................................................61
4.10 2-D PCA Overall Relative Deformation.........................................62
IX


CHAPTER I
INTRODUCTION
In the northern highlands of Tanzania, nestled within an uplift of the East African Rift system, and surrounded by innumerable eroding fossiliferous deposits and outcrops, is one of the most important and spectacular paleoanthropological finds in the world: the Laetoli Site G hominin footprint trails. Imprinted in volcanic ash 3.66 million years ago by Pliocene hominins, the Site G trackway, discovered in 1978 by a team led by Mary Leakey, upended the scientific community, and was a crucial advance in our understanding of human evolution (Leakey and Hays 1987). In addition to their remarkable state of preservation, the footprints confirmed the deep roots of bipedality in hominin evolution, and their association with the then recently described Australopithecus afarensis demonstrated the emergence of the suite of bipedal locomotor adaptations before encephalization in the evolution of our clade. The trackways have remained a continuing source of important research and controversy, even four decades later. The three track-makers' locomotor capabilities, gait, body size, pedal morphology, social grouping, and paleohabitat have all been intensely scrutinized and debated, and the continued lack of hominin pedal fossils found at Laetoli casts a shadow over many analyses. The lack of any almost complete associated foot elements at Hadar, where the vast majority of Au. afarensis specimens have been discovered, is similarly frustrating. While scarcely a season has passed without field research at the 20km2 Laetoli site at large, and despite numerous erosional and excavated exposures of associated animal trackways, it wasn't until 2015 that additional hominin prints were discovered. Preserved in the same geological horizon and located only 150 meters to the southwest of the G trackway (Masao, et al. 2016), the markedly larger Site S footprints have the potential to reinvigorate hominin locomotor research and stoke new questions about the hominins the inhabited the Pliocene Laetoli landscape.
1


While debates regarding speciation within the hominin clade and intra-specific variation have long been a part of the paleoanthropological dialogue, in the past two decades researchers have to contend with both an increasingly complex and diverse hominin evolutionary tree, particularly the marked morphological variation in areas such as locomotor repertoire, body proportions, and masticatory complexes.The discovery of the Site S footprints provides an important illustration of the emerging picture of hominin variation, both within and between species. With this in mind, in order to situate these newly discovered footprints within (or against) our current understandings of hominin evolution and behavior, comparative analyses must be undertaken with other, similar datasets, including the Site G trackway and habitually unshod modern humans. While there are numerous taphonomic, substrate, and biomechanical variables to consider in footprint formation, they provide us with the best direct evidence of how hominins moved, and any notable differences that can be discerned between different populations is potentially illuminating of morphological and/or locomotor variation.
Some specific aims of this thesis include:
1. Direct comparison, qualitatively and quantitatively, of the three currently known sections of the Site S trackway with the Site G trackway via use of 3-D photogrammetry.
2. Alongside the direct visual qualitative and quantitative comparison of these two sets of trackways, additional comparison with unshod modern human data utilizing footprint measurement and gait metric data, and a discriminant function analysis.
3. Determine if the footprint morphology of the Site S prints demonstrates significant differences from those of Site G individuals via Procrustes fitted landmark principal components analysis.
2


CHAPTER II
BACKGROUND RESEARCH AND CURRENT LITERATURE Theoretical Background
In studies such as this, it is important to contextualize the research within our current understanding of the evolution of bipedality in the hominin clade and within models of hominin behavior and mobility. One of the longest standing hypotheses regarding the evolution of upright posture and bipedal locomotion posits that it emerged as an adaptation to the opening of the East African landscape from primarily forested to savannah grasslands. However, there is a growing amount of evidence from early hominin sites that indicate many of the earliest-known definitive bipeds were inhabiting environments that were still largely wooded or mosaic (Potts 1998; Behrensmeyer and Reed 2013). The existence of early putative Miocene hominin Ardipithecus in what is argued to be a closed woodland environment, and Sahelanthropus in what is interpreted as a more open habitat demonstrates the variable nature of early hominin paleohabitats (Lovejoy, 2009; Brunet, et al. 2002). Many of the models, such as Wheeler's (1991) thermal regulation hypothesis, that centered around open grasslands as the driving force of bipedality, appear to be teleological and concerned with function over adaptation and behavioral response. Nevertheless, there is a notable expansion of African grasslands within the Pliocene, and as Dominiguez-Rodrigo (2014) points out, savannah as a biome encompasses a variety of habitats, including closed woodland, which entail a variety of adaptive responses. The mosaic nature of many Miocene and early Pliocene African environments largely precludes open grasslands as the driving force of bipedality, as does climatic evidence that shows aridification and a marked increase in continent-wide grassland expansion not occurring until later in the Pliocene, around 3.0-2.8 million years ago (Behrensmeyer 2006). Thus, other selective pressures had to be acting upon the earliest hominins to push bipedality as
3


an adaptive response after its initial emergence in the Miocene. Behavioral and adaptive flexibility in the face of climatic and environmental uncertainty is often posited, which also ties into diet and subsistence behavior, and cannot be extracted from discussions regarding mobility (Dominiguez-Rodrigo 2014; Kuhn, et al. 2016). While external ecological factors are crucial to understanding hominin adaptation, their behavioral responses, both on an individual and on a population level are also necessary to understand.
Behavioral ecology models have a complicated relationship with anthropology, yet they remain the best framework with which to understand early hominin behavior and mobility, particularly prior to the earliest archaeological evidence. In order to understand what drove bipedality as an adaptive response, regardless of the specific ecological context, one needs to approach what drives hominin mobility from a theoretical perspective. Nathan, et al. (2008), in their movement ecology framework, refer to this as the "internal state" of movement. While constrained by the capabilities of the organism itself and the external factors, the internal state of movement cuts across theoretical models based on behavioral ecology. This relationship between internal and external dynamics shapes mobility throughout human evolution, from the australopithecines to Paleolithic hunter-gatherers and onward. One of the most basic driving urges is subsistence, and as Kuhn, et al. (2016) state, "variation in diet directly implicates movement." For early hominins, habitat and dietary variation can theoretically be correlated to differences in subsistence behavior, and potentially, locomotor adaptation, which in turn, could be linked to adaptive radiation. The variable presence and retention of arboreal traits in australopithecines well into the Pleistocene and the emergence of Homo suggest behavioral variability reflecting adaptations to different habitats (Ward 2002). While some models of the evolution of bipedality have explicitly linked its emergence to subsistence and provisioning behaviors (Lovejoy 1981; Hunt 1994), it is likely that dietary variables were coupled with other behavioral changes and adaptations alongside the evolution of bipedal locomotion.
4


With both extant primate and modern human behavioral models, the issues with inferring early hominin behavioral ecology stem from the general problems with the use of analogy within the paleontological and archaeological record. Behavior does not necessarily preserve directly in either record, so researchers must utilize the evidence at their disposal, namely skeletal morphologies, artifacts, and associated faunal remains, to infer it indirectly. For early hominins, the issue lays in the fine line between over-anthropomorphizing hominin behavior and drawing too much from the behavior of our closest evolutionary relatives, who continue to occupy the traditional niche of apes, and have undergone their own evolutionary trajectory. In both instances, despite the issues, there is merit in some of the perspectives that behavioral ecology can offer studies of early hominin mobility. Humans and early hominins alike, are nevertheless primates, and extant species, and Pan in particular, offer the easiest model of social behavior for comparison (Foley and Gamble 2009). While other animals may practice coordinated hunting, food sharing, tool use, etc. (Sayers and Lovejoy 2008), other primates and all the different ecological and social adaptations they possess, are the best comparative model currently at our disposal. This is true with not only adaptive behavior, but also with morphological adaption and radiation. This is best exemplified by the frequent existence of more than one primate species in a given habitat, which is seen in both the fossil record and in modern day primate communities.
5


Current Literature
Geology and Paleoecology of Laetoli
The geology of the Laetoli area has been extensively mapped and studied, especially since extensive research began at the site in the 1970s. The Plio-Pleistocene sedimentary and volcanic deposits (Fig. 2.1), which lay directly upon Precambrian metamorphic basement rocks, consist of a series of beds spanning from 4.36-2.3ma: Lower Laetolil (~4.36ma 3.85ma), Upper Laetolil (~3.85ma 3.63 ma), Ndolanya (-3.58 -2.66ma), and the Naibadad Beds (-2.15ma 2.06ma). The Upper Laetolil beds and the younger Ndolanya beds are separated by a distinctive yellow marker tuff throughout much of the site, likewise with the Ogal lavas separating the Ndolanya Beds from the upper deposits. The Upper Laetolil bed, the largest and most fossiliferous of the deposits, and the sequence that contains the footprint tuff, is largely comprised of aeolian and air-fall/fall-out tuffs, which, due to their geographic extent at the Laetoli site, have undergone extensive, but variable, fluvial and standing water reworking at the different localities (Hay 1987; Musiba, et al.
2007; Ditchfield and Harrison 2011). The major fossiliferous tuffs (Tuff 6 through Tuff 8, largely) are dominated by volcanic ash fall-out deposits. Tuff 7, of which the lower portion has been dubbed the footprint tuff, is the thickest and most distinctive (Ditchfield and Harrison 2011). The footprint tuff consists of fine-grained, heavily compacted carbonatite ash, which varies in thickness from 12-15cm, and is subdivided into two differing lithological layers, which in turn document numerous separate eruption and ash fall events, many of which contain the eponymous footprints (Hay 1987). The frequency of the layers, and the excellent preservation of the footprints and very existence and high preservation of raindrop imprints, points towards very short depositional periods, likely only a few weeks (Hay 1987).
6


Figure 2.1: Stratigraphic sequence at Laetoli (left) (Harrison 2011) and microstratigraphy of
the footprint tuff (right) (Hay 1987).
This has significant implications for studying the paleoecological makeup of the Laetoli site during this period, an opportunity that most time-averaged fossil deposits cannot provide. The greater Laetoli/Eyasi Plateau region (Fig. 2.2) is dotted with ancient volcanoes that are responsible for the ash fall tuffs, although it is generally believed that the volcano Satiman is the likely source of the ash fall deposits making up the footprint tuff, based on its age and the geochemical composition of the rocks, sediments, and lava flows found at the extinct volcano (Hay 1987; Meldrum et al. 2011). Although it is not a definitive relationship, as depositional and temporal factors have resulted in numerous geologic discrepancies, (Meldrum, et al. 2011; Zaitsev, et al. 2011), Satiman currently remains the most plausible
7


candidate for creation of the Laetolil beds.
The preservation of the animal trackways in the ash-fall tuffs is a crucial piece of evidence for the paleoecological construction of Pliocene Laetoli, given the relative dearth of fossilized remains at the site (Su and Harrison 2008). The discovery of the Site G hominin trackway and its central role in understanding the evolution of bipedality has given constructing the paleoecology of Pliocene Laetoli habitats accurately an added importance. In the initial monograph published on the site, Leakey (1987) argued for the paleoecological interpretation of a largely dry, open-grassland biome. This interpretation was supported by their analysis of the faunal evidence accumulated at the time, and possibly influenced by the then prevailing "open savannah" hypothesis for the emergence of hominin bipedality. Since the initial publications of research emerging from the Laetoli site, there has consistently been competing interpretations of the paleoecological evidence, with taphonomic and sampling biases argued to be the culprit for earlier "Serengeti-like" interpretations (Musiba, et al. 2007). There is mounting evidence for a greater occurrence of both closed and open wooded habitats, a trend that, as noted above, is becoming broadly seen across many late-Miocene and Pliocene hominin sites (Behrensmeyer and Reed 2013). Faunal, paleofloral, dietary isotopic, and pollen evidence accumulated over the last three decades of research at Laetoli support a largely wooded landscape with a mix of more open grassland habitats (Kingston and Harrison 2007; Harrison 2011; Su 2011).
Other recent interpretations agree with this general assessment, but argue that the paleoenvironment of Pliocene Laetoli was much more variable and mosaic in its habitat makeup, likely containing starker divisions in its biomes than seen presently, due to highly variable local climatic factors (Musiba, et al. 2007; Andrews and Bamford 2008; Musiba, et al. 2008). Analyses focusing on the past and current vegetation landscape, including fossil remains and isotopic signals at Laetoli note many similar biomes, but a distinct difference in the overall make-up at the site over time. Both Andrews and Bamford (2008) and
8


1 Volcano
^T Paleontological / Archaeological site
... Approximate contour of paleo-lake Olduvai 1.75 million years ago
Figure 2.2: The Northern Highlands of Tanzania, showing paleo-volcanoes and notable sites. Volcano (1) is Satiman. From Barboni (2014).
9


Figure 2.3: Map of the Laetoli region at large, from Harrison (2011).
Figure 2.4: Map of the Laetoli Site. Localities are designated by numbers, footprint sites by
letters. Modified from Musiba, etal. (2008).
10


Barboni (2014) note the existence of C4 grasslands in Pliocene Laetoli, yet argue that they only make up a smaller portion of the vegetative biomass at the site compared to C3 plant material. However, Barboni (2014) notes that in the phytolith evidence, a steady expansion in C4 grasses relative to woody cover can be seen after the deposition of Tuff 6 (~3.7Ma). Skewing towards the more wooded end of the biome spectrum, Pliocene Laetoli would likely have been a patchwork of woodland, shrub lands, and grasslands of varying compositions, and more heavily wooded than the present day Laetoli region (Harrison 2011). This has important implications for subsequent interpretations of the locomotor adaptations of the Laetoli hominins, as well as other crucibles of bipedal evolution in Africa.
The effects of the volcanic and tectonic activity on the landscape and paleoecological makeup at Laetoli during the mid-Pliocene also cannot be discounted. Both Andrews and Bamford (2008) and Barboni (2014) both note how the processes of uplift resulting in the creation of the East African rift would have created variable rainfall in Pliocene Laetoli, which arguably contributed to a stark mosaic-like habitat structure. The frequent disruptions from volcanic activity and ash-fall undoubtedly affected both the floral and faunal communities, although to what extent is unclear. Andrews and Bamford (2008) argue that much as present volcanic and tectonic activity affects the vegetation structure on either side of the East African Rift, it would be have particularly notable in the Pliocene, contributing to habitat variability. Ditchfield and Harrison (2011) argue that the effects of the periodic ash-falls would be incredibly disruptive for woodlands, resulting in periods with higher amounts of open-grasslands. However, they argue that these disruptions would be relatively short-lived, and the paleoenvironmental balance of Pliocene Laetoli would rebound as soils redeveloped to allow for woody growth. A frequently shifting landscape and vegetation structure seems to be the ecological settings in which the Laetoli hominins inhabited, and it likely had an effect on their adaptive responses.
11


Locomotion of the Laetoli Track-makers
There are several notable features that separate the morphology of the modern human foot from that of the rest of the apes, all of which are functionally important in our obligate bipedal locomotion. Alongside our comparatively shortened toes, which decrease in size from the first digit, one of the most noticeable features is our robust, non-grasping hallux. The reorganization of the forefoot allows it to function as an effective lever during bipedal gait, with the robust hallux playing a key role in propelling the body forward. In addition to their abductable hallux, non-human primates also lack the derived medial longitudinal arch, possessing only the transverse arch of the foot. This gives their feet (and as a result, their footprints) a flattened, splayed appearance, with a noticeably large gap between their first and second digits. The medial longitudinal arch is a significant evolutionary development for obligate bipedalism, allowing for greater stability, shock absorption, and medial weight transfer (Aiello and Dean 1990). While arch formation and structure is variable in modern humans, it remains a key difference separating human-like foot function from less-derived forms, even among modern humans with non-pathological flat feet (Klenerman and Wood 2008). Other key shifts in modern foot construction include reorganization of the ankle joint and a relatively stiff mid-foot, as arboreal-adapted apes have a great deal of mid-foot flexibility, known as the mid-tarsal break (Aiello and Dean 1990). While there are limits to the inferences that can made on footprints, they nevertheless can provide insights into certain skeletal and soft tissue morphologies and gait patterns by looking at their construction and pressure distribution (Bennet and Morse 2014). However, as the Laetoli prints have demonstrated, interpreting these subtle signals can be contentious.
The Site G footprints are undeniably made by a group of habitually upright bipedal individuals, a fact recognized immediately upon their discovery (Leakey and Hay 1979), However, debates within the paleoanthropological community have largely centered around
12


the relative human-like nature of the bipedality seen in the footprints, as well as what hominin species were responsible for their creation. Upon their discovery in 1978, only two individuals were initially recognized (Leakey and Hay 1979), although as additional, better preserved prints were discovered, it became clear that the second, larger, tracks were in fact the prints of a third individual overlaying those of the second (Leakey 1987). More recent analysis of the overlaying prints has provided glimpse of a potential fourth individual in G-2/3, based on the presence of what appears to be additional hallucal imprints (Bennet, et al. 2016a). However, this could also be a false signal from the trackway cast, as White and Suwa (1987) remark that lagomorph prints are abundant in the footprint tuff, and note the presence of a track crossing the G 1-36 print. The overlaying of these two (or three) individuals has complicated analysis and comparison of the two parallel trackways, as the prints of the G-3 individual largely obliterated the posterior portion of the G-2 prints, preventing accurate measurement of the true length of the G-2 individual's footprints (Tuttle 1987). While some prints of the G-2/3 trackway allow for better estimation of the shape and metrics of the G-2 individual, most researchers have opted to focus their analyses on G-1 and G-3, the two smaller of the three individuals.
While preservation of the prints varies wildly along the trackways (particularly in the northern end), the G-1 prints themselves are noticeably better preserved than the G-2/3 trail, so much that Leakey (1979) described them as likely having been imprinted on a more dense, compacted surface compared to the other trail, suggesting that the individuals walked through the area at different times within the same ash-fall event. However, White and Suwa (1987) remarked that the different walking surface hypothesis had been largely abandoned by the original researchers after further study of the site. Additionally, in their analyses of the Site G Trackway, both Robbins (1987) and Tuttle (1987) remark that it is likely that all three individuals were walking through concurrently, and the G-1 and G-3 individual were attempting to keep pace with the G-2 individual, based on how closely the
13


prints mirror each other in placement. Tuttle (1990), nevertheless states that temporally separate footprint creation events are still likely and cannot be definitively fully ruled out.
In addition to arguing for the presence of a medial longitudinal arch, both Robbins (1987) and Tuttle (1987) remark on the affinities the tracks share with modern human locomotor function, particularly in the heel, ball, and toes. Weight transfer from heel to toe along the lateral side of the arch, followed by a medial shift towards the adducted hallux and its notably deep impression, suggesting toe-off prior to swing phase, is noted by both authors. Tuttle (1987) goes as far to say that the prints are "indistinguishable from those of modern humans," not noting any phalangeal curvature or hallucal gaps outside of modern human variation, although their stride lengths were noticeably shorter than seen in modern humans. White and Suwa (1987) dispute this observation by Tuttle (1987), but nevertheless state that they are significantly more similar to modern human footprints than any other known primate. Most importantly, they state that the foot of the track-makers need not be identical to that of H. sapiens to perform human-like bipedal locomotion. In their comparative analysis utilizing footprints of the habitually unshod Machiguenga peoples, Tuttle, et al. (1990) reaffirm their argument as to the remarkable human-like nature of the prints. Stern and Susman (1983) disputed the assertions of Leakey (1987), Robbins (1987), Tuttle (1987), and White and Suwa (1987), arguing that the Laetoli footprints resemble a much more transitional form, sharing affinities with chimpanzee footprints and displaying poorly developed morphologies that are present in modern humans. Tuttle, et al. (1990) aggressively disputed the assertions made by Stern and Susman (1983), claiming that they were built upon incomplete and incorrect data and analysis. Further comparative studies with the Hadzabe of Tanzania by Musiba, et al. (1997) bolstered the argument that the Laetoli prints were remarkably human-like in form and function. Given the differences in stature, walking speed and stride length between the Hadza and Machiguenga peoples (Musiba, et al. 1997), the affinities seen in the Laetoli prints are notable.
14


Nevertheless, the debates regarding the nature of the Laetoli track-makers has continued into the 21st century. A more recent statistical study by Berge, et al. (2006) reported some affinities with chimpanzee gait and morphology, such as foot roll-off at stance phase and pedal proportions. However, their analysis overwhelmingly supports the notion that the Laetoli track-makers displayed more human-like traits in their pedal function than those of a chimpanzee, particularly in the anterior part of the foot (Berge, et al. 2006). Raichlen, et al. (2008) studied the locomotor biomechanics of the Laetoli prints, and determined that the track-makers could have either used modern human-like extended hind limb gait or more chimpanzee-like "bent knee-bent hip" gait to create the footprints.
However, in a later study utilizing three-dimensional laser-scan modeling of pressure kinematics, Raichlen, et al. (2010) determined that the Laetoli track-makers walked with extended limb gait, as the prints displayed significantly more morphological similarities with experimental modern human prints walking normally than those made by modern humans walking with bent knee-bent hip gait. A pedobarographic statistical mapping analysis conducted by Crompton, et al. (2012) reached similar conclusions regarding the human-like locomotion of the Laetoli track-makers.
Bennet, et al. (2016b) and Hatala, et al. (2016) most recently have offered competing conclusions regarding the function of the Laetoli track-makers feet. Whereas Bennet, et al. (2016b) maintain the belief that the Laetoli tracks are functionally indistinguishable from modern humans, and promote the idea that hominin foot functionality has possibly remained in evolutionary homeostasis for the last 3.66Ma, Hatala, et al. (2016) argue that the Laetoli tracks demonstrate a form of locomotor mechanics unlike humans or chimpanzees, largely based on differences in medial foot rotation and knee and hip flexion during gait. Meldrum, et al. (2011) similarly distinguish the Laetoli track-makers from modern human morphology, particularly with what they interpret as evidence of a midfoot flexibility. The mid-tarsal break, which is the ability to lift the heel independently from the rest of the foot, is often interpreted
15


as a primitive condition absent in modern humans, and prominent in other primates, including chimpanzees (DeSilva 2010; Meldrum, et al. 2011). Greater midfoot flexibility results in noticeably different morphological signals in footprint formation, particularly in pressure dynamics. Meldrum, et al. (2011) interpret several of the Site G tracks as displaying evidence of greater midfoot flexibility, as well as the lack of a medial longitudinal arch. While they do not dispute the clear adapations for terrestrial bipedality seen in the Site G footprints, such as the shortened toes and the adducted hallux, Meldrum, et al. (2011) assert that the Laetoli track-makers' midfoot flexibility distinguishes them from modern human locomotion. However, Meldrum, et al.'s (2011) interpretation is based on the assumption that the mid-tarsal break occurs at the calcaneocuboid joint, which correlates with skeletal reconstruction overlays on the G-1 prints. DeSilva (2010) disputes this long-held interpretation of the mid-tarsal breaks, and argues for its placement more anteriorly, at the cuboid-metatarsal joint. Additionally, DeSilva, et al. (2015) argue that midfoot flexibility is present in varying degrees in modern humans, and is much more complex of a feature than previously realized, and does not necessarily disprove the existence of a medial longitudinal arch. Bennet and Morse (2014) also dispute the mid-tarsal break interpretation, and posit substrate and gait mechanic factors for footprint morphologies argued by Meldrum, et al. (2011).
Locomotion of Australopithecus afarensis
Deeply intertwined with the debate regarding the functional signals interpreted from the Site G trails is whether Au. afarensis is the species responsible for creating the footprints. The only presently known fossilized hominin remains at Laetoli contemporaneous with the trackways are assigned to Au. afarensis, yet the collection is relatively diminutive compared to the amount of material from Hadar, Ethiopia assigned to the Au. afarensis hypodigm. Furthermore, there have not yet been any hominin pedal remains recovered
16


from Laetoli, while the Hadar site posses nearly the entirety of the Au. afarensis pedal collection, none of which comprises a complete enough foot to make direct inferences on the Laetoli tracks. Taphonomic factors can be held accountable, as pedal fossils are rare, particularly more complete specimens such as the OH 8 foot, for a variety of reasons, not the least of which is their small size and likelihood of preservability. At Laetoli, the taphonomic conditions and paleoenvironment have resulted in a generally smaller fossil collection compared to many fossil localities such as Hadar or nearby Olduvai Gorge, likely due to its lack of permanent lacustrine or riverine water sources (Harrison 2011). Hominins are a rarity amongst fossil assemblages, but it is particularly so at Laetoli, which may the result of the aforementioned taphonomic bias, or potentially habitat preferences of the species (Su and Harrison 2008).
While numerous pedal elements attributed to Au. afarensis, dated from 3.4Ma -3.0Ma, have been discovered at Hadar, particularly from the bountiful AL-333 locality, there has only been one associated assemblage of foot elements, AL 333-115 (Fig. 2.5). This
17


largely fragmentary collection consists of five proximal phalanges, two middle phalanges, and five metatarsal heads, which while informative, lacks the more robust elements that typically preserve and can be used for comparative analysis (White and Suwa 1987). Other than the small collection from AL 288-1 ("Lucy"), consisting of a talus and two phalanx fragments, the rest of the Hadar pedal collection consists of isolated specimens (Johanson, etal. 1982; Ward, etal. 2012).
Taken in full, numerous similarities can be seen across the combined Hadar collection, as well as a few key differences in the tarsal specimens. The metatarsals, while still curved, are only slightly more so than modern humans, and display distinct dorsal doming of the heads, which separates them from the condition seen in extant apes. The dorsoplantarly expanded bases and sulcus separating the metatarsal heads from the epicondyles resembles the morphology of modern human metatarsals (Ward, et al. 2012). The phalanges are longer and more curved than in Homo, but less so than in modern apes, and comparable to Ar. ramidus. Similarly, the metatarsal articular facets of the phalanges are more dorsally inclined, at the lower end of what is seen in modern human variation, which is argued to reflect extension during the terminal stance phase of bipedal gait (Stern and Susman 1983; Ward, et al. 2012).
The navicular of Au. afarensis has been a particularly controversial element, because of its large tuberosity and flattened appearance, a primitive condition which has been argued as evidence for its role in weight-bearing during gait, which would possibly presuppose the existence of a medial longitudinal arch (Stern and Susman 1983). However, in addition to the expanded knowledge of metatarsal morphologies that have resulted from continued recovery at Hadar, newer navicular specimens show an expanded area for the insertion of the plantar cubonavicular ligament and a groove for calcaneonavicular ligaments, as seen in modern humans, which challenges this interpretation (Ward, et al. 2012).
18


The calcaneal specimens have been similarly controversial, due to their importance in the striking phase of bipedal gait. While the calcanei attributed to Au. afarensis retain the oval-shaped appearance seen in extant apes as well as the "massive" peroneal trochlea and flattened posterior talar surface, they are substantially increased in robusticity in comparison, and are markedly more human-like in both cross section and in tuberosity volume (Latimer and Lovejoy 1989). Additionally, Latimer and Lovejoy (1989) disagree with the assessment of Stern and Susman (1983), and argue for the presence of a lateral plantar process in the Au. afarensis calcanei, which is a distinct trait seen in later hominins and modern humans.
The tali of Au. afarensis have also been a point of contention, as there are marked differences between that of Lucy and the larger specimens, such as AL 333-147. Previous analyses pointed to the more primitive aspects of Lucy's talus, such as the tightly curved, deeply grooved trochlea, and the anterior extension of the fibular malleolar facet on the lateral side of the talar neck, as evidence of arboreal retentions in the species (Susman and Stern 1983). However, the more recently discovered larger specimens lack these morphologies, and appear to be in the lower range of modern human variation. These tali exhibit all the morphologies associated with habitual bipedality, including a vertically oriented shank and groove for the flexor hallucus longus muscle, and a human-like axis of rotation at the talocrural joint (Ward, et al. 2012). The differences seen between several of the pedal specimens attributed to Au. afarensis have been attributed by several of the researchers at Hadar to differences in body size and sexual dimorphism (White and Suwa 1987; Ward, et al. 2012), although the potential for different modes of bipedality is often alluded to (Stern and Susman 1983; Harcourt-Smith and Aiello 2004). The composite nature of most analyses regarding the foot of Au. afarensis has exacerbated this issue
While few researchers now dispute the terrestrial bipedality of Au. afarensis, reconciling the morphology of the Site G footprints with the skeletal anatomy seen in the
19


Hadar pedal assemblage has prompted debate amongst researchers (Kimbel and Delezene 2009). Tuttle (1987; 2008) has maintained that the morphologies exhibited by the Hadar specimens of Au. afarensis would exclude the species from having made the Site G footprints. In addition to the potentially weight-bearing naviculars, which could contradict the evidence of a medial longitudinal arch in the Site G prints, Tuttle, et al. (1990; 1991a), focus on the disconnect in phalangeal curvature. Whereas the Site G prints display shortened toes statistically indistinguishable from a Machiguenga footprint sample, and faint lateral toe impressions, Tuttle, et al. (1990; 1991a) argue that the length and level of curvature seen in the Hadar phalanges would have prevented them from making such footprints. Harcourt-Smith and Aiello (2004) also argue that the enlarged tuberosity of the Au. afarensis navicular is evidence of its weight-bearing nature, and likely a lack of a human-like arch. Many of the Hadar researchers maintain that the morphologies of the pedal elements, while retaining some primitive morphologies in certain areas, are nevertheless capable of having created the prints at Laetoli (White and Suwa 1987; Ward 2002). To support their assertions, White and Suwa (1987) reconstructed an Au. afarensis foot scaled to the size of the G-1 tracks from a composite of AL 333 elements and the OH 8 midfoot. Stating that their reconstruction does not definitively prove that Au. afarensis was the track-maker, they argue that the species nevertheless cannot be excluded. Ward (2002; 2013), in addition to agreeing that Au. afarensis was the likely track-maker at Laetoli, asserts that the body plan and lower limb function of australopithecines emerged within the Au. anamensis/afarensis lineage and was maintained without much variation for at least 700,000 years. Alternatively, Stern and Susman (1983) argued that the transitional morphology of the Au. afarensis foot matched well with the transitional morphology they argued to see in the Laetoli trackways. Tuttle (1990) disputed the arguments made by White and Suwa (1987), specifically citing their use of the OH 8 foot, which at roughly 1,75ma, is significantly younger than any of the Hadar Au. afarensis specimens. DeSilva's (2010) argument in favor of a stable midfoot in
20


early hominins, Au. afarensis included, contradicts Meldrum, et al.'s (2011) interpretation of a mid-tarsal break in the Site G prints, which also utilized the reconstructed foot from White and Suwa (1987). Depending on one's interpretation of the morphologies seen in the Au. afarensis pedal specimens, either scenario regarding midfoot flexibility could be subsumed into the species.
A great deal of the debates regarding the locomotion of the Laetoli track-makers and Australopithecus can be subsumed into broader discussions of the relationship between form and function of extinct species. This is particularly true in footprints, which are ultimately imprinted by soft tissue, but the functional aspects of the pedal bones play a key role in their creation. The issues of sorting out primitive, adaptively-neutral, retentions versus traits that are actively maintained selectively and behaviorally is particularly difficult with early hominins. Skeletal adaptations in the foot play a key role in bipedal locomotion, but the interplay between morphological adaptation and locomotor function extends up through the entire body, in both hard and soft tissue. Researchers, depending on whether their perspective focuses on selective history or actual functional abilities, can interpret the total body plan in strikingly different ways (Ward 2002). While the morphological form of fossilized remains can provide some clues as to its function, caution needs to be taken, particularly when making inferences on adaptive significance. As Ward (2002) notes, there are numerous examples in the fossil record of a disconnect between form and function, particularly in primate dentition, although this could easily extend into the postcrania. To what extent any individual morphological differences suggest behavioral and functional differences, and if this comprises intraspecific variation or species-level distinction is another theoretical and interpretive point of discussion.
21


Pliocene Hominin Adaptive Radiation and Australopithecine Locomotor Variation
While Au. afarensis is the only currently known species at Laetoli contemporaneous with the footprint tuff, Tuttle (1987; 1990; 1991b; 2008) maintains the possibility of another, yet unknown, species of hominin may be responsible for creating the Site G trackway. Locomotor and morphological debates surrounding Au. afarensis aside, our expanding knowledge of Pliocene hominin diversity has made this notion increasingly plausible. Speciation and the amount of accepted intraspecies morphological variability is a continuously debated issue within studies of human evolution. Au. afarensis, as a commonly accepted hypodigm, spans an enormous geographic range across East Africa, and a substantial temporal range of roughly 700,000 years, potentially longer if its anagenetic lineage with Au. anamensis is accepted (Kimbel, et al. 2006). The far reaching and long-lived success of this hominin could be interpreted as it being an adaptive generalist, both in body plan and in behavior, with morphological variability within the taxa explained by sexual dimorphism and temporally progressive evolutionary change (Kimbel, et al. 2009; Ward 2013). However, given the discovery and broadening acceptance of additional contemporaneous hominin species, including Kenyanthropus platyops, Au. bahrelghazali, and Au. deyiremeda (Brunet, et al. 1996; Leakey, et al. 2001; Haile-Selassie, et al. 2015), as well as the recently discovered Burtele foot (Fig. 2.3), which differs markedly from composite feet of Au. afarensis (Haile-Selassie, et al. 2012), it is likely that there was much greater diversity within the hominin clade in the Pliocene than previously interpreted. The interpretation of Au. afarensis as an adaptive generalist with varying morphologies may in fact mask the existence of more than one species at sites such as Hadar and Laetoli.
The recently discovered Burtele foot (Haile-Selassie, et al. 2012) (Fig. 2.6), provides direct evidence of morphological variability among hominins prior to 3.0ma. Dated at 3.4ma, BRT-VP 2/73 from the Woranso-Mille study area in Ethiopia is contemporaneous to Au. afarensis at nearby Hadar. Consisting of a small associated collection of metatarsals and
22


phalanges, the foot is directly comparable to specimens such as AL 333-115. Despite sharing certain bipedally-adapted features with australopithecines, there are numerous unique aspects to this specimen. Its potentially adductable hallux is particularly perplexing, given its age. The base of Mt I is tall and deeply concave, lacks the dorsal doming of the head seen in australopiths, and is short relative to the other metatarsals. This ratio puts
I
Figure 2.6: BRT-VT-2/73 partial foot from Burtele, Ethiopia.
From Haile-Selassie, etal. (2012).
the Mt I of BRT-VP 2/73 in the morphological range of extant apes, and as such, was likely not used for toe-off. The second ray, which consists of the Mt II and proximal and intermediate phalanges, further evidences the abducted nature of the hallux. Its torsion towards the hallux is less than seen in extant apes, but is significantly more than that of modern humans (Haile-Selassie, et al. 2012). The fourth ray is unique in that its metatarsal lacks the expanded stabilizing base morphology seen in later hominins, including Au. afarensis (Ward, et al. 2012), but also in its longer length than both Mt I and Mt II, which is a unique, potentially derived, condition not seen in any extant primates. All of the phalanges of the Burtele foot exhibit strong curvature, although less than in Pan, and generally resemble
23


those of Ar. ramidus and Au. afarensis (Haile-Selassie, et al. 2012). In essence, this specimen appears to continue the pattern seen in Ar. ramidus, but with continued refinements for bipedality, and stands in contrast to Au. afarensis.
In addition to the Burtele foot and its clear morphological differences from other reconstructed australopithecine feet, there is growing evidence, particularly from South African sites, for the morphological variability within the australopithecines in the Plio-Pleistocene. While Au. africanus has been argued to continue the same basic locomotor pattern as Au. afarensis (Ward 2013), the Stw 573 ("Little Foot") skeleton, dated at 3.5Ma, has been argued by its discoverers to not only constitute another species (Au. prometheus), but also to display noticeable morphological differences (Clarke 2013). While most of the joint morphology and orientation does not differ greatly from A. afarensis or even H. habilis, the hallux is argued to be slightly divergent, which would suggest abductability. Additionally, the cuboid-calcaneal articular surface is bowl-shaped, which in addition to the more archaic morphologies of the lateral cuneiforms, points towards greater mid-foot mobility. However, the non-weight bearing navicular suggests that Stw 573 had the presence of somewhat of a longitudinal arch (Clarke 2013).
Au sediba, although geologically much younger than most known australopithecines, dated at 1.98Ma, similarly has a mix of primitive and derived traits in its locomotor complex, including its feet (Zipfel, et al. 2011). The talus exhibits some human-like morphologies, but the neck angle and torsion angle of large talus head are both argued to be more ape-like, as is the lateral plantar process, which is positioned more superiorly, not in a weight bearing position. This would potentially result in elevated heel stress, especially in light of its more mobile calcaneocuboid joint, relative to modern humans. Its calcaneal tuber is significantly more gracile than A. afarensis and modern humans, but contains an attachment for a long plantar ligament, which stabilizes the human mid-foot. The calcaneus overall resembles the ape-like morphological condition moreso than that of modern humans. However, the
24


metatarsals largely resemble humans in morphology, and the positioning of the talus strongly suggests the presence of an arch. This odd mix of traits, particularly the positioning of the lateral plantar process, has led some researchers to suggest that A. sediba practiced a form of bipedalism unique from both A. afarensis and modern humans (Zipfel, et. al 2011). A recent study of calcaneal robusticity amongst australopithecines demonstrated marked differences between certain species, which was argued to demonstrate differences in mobility and habitat exploitation (Prang 2015).
The taphonomic bias of the fossil record towards craniodental remains has led to further bias in studies of early hominin variability and diversity, and the small size of the australopithecine postcranial collection has limited researchers to some extent (Ward 2013). However, if researchers are beginning to accept Pliocene hominin dietary diversity connected to craniodental morphological variation, it is not a stretch for morphological variability connected to the locomotor suite of adaptations to be similarly accounted for in the fossil record (Haile-Selassie, 2016).
Ecological context plays a significant role in driving adaptation, particularly as it concerns mobility, diet, and subsistence behavior (Potts 2007). Given the interconnected nature of these aspects, a high degree of environmental variability, on both the regional and local-habitat scale, would provide impetus for an adaptive radiation in both cranial and locomotor morphology. The degree of environmental variability in Pliocene Africa is well documented from faunal, ocean core, and geologic evidence, as is the existence of australopithecines in a wide range of paleohabitats (Behrensmeyer and Reed 2013). The degree to which hominin adaptation to habitat and environmental variability could result in speciation, particularly as it concerns Au. afarensis, is nevertheless a subject of debate. Behrensmeyer and Reed (2013) argue that as a genus, australopithecines were remarkably adept at exploiting different environments, and leave open the possibility of multiple hominin species co-existing at sites, despite maintaining the interpretation that the specimens from
25


Hadar, Laetoli, and Woranso-Mille are all Au. afarensis. Foley (2002) disputes the adaptive radiation hypothesis for australopithecines, and instead posits a dispersal-based divergence of early hominin species. However, this is largely based on broad time-averaged geographic evidence, as well as the "course-grained" nature of the pre-Homo species record. While dispersal and patterns of mobility among hominins no doubt influenced adaptation, gaps in the fossil record do not preclude habitat-level radiation, which does not need to be "explosive" in nature or require intense specialization to allow for speciation (Foley 2002). Given the relatively novel, evolutionarily speaking, nature of hominins' locomotor capabilities and the environmental variability of Pliocene east Africa, adaptive experimentation can reasonably be assumed to lead to some degree of regional speciation and admixture (Potts 1998). Additionally, if high mobility can be expected of early hominins, which the existence of Au. bahrelghazali in Chad points to, it is reasonable to expect populations to have overlap in their dispersals. As Haile-Selassie (2016) points out, modern chimpanzees and gorillas are adaptable to broad variety of niches, and often can overlap in their ranges and exploited environments. At Olduvai Gorge, both robust and gracile hominins existed side by side, and appear to have exploited different ecological niches (Barboni 2014), and it is likely that similar ecological divergence occurred amongst earlier hominins at Pliocene sites, Laetoli included.
However, the taxonomic validity of many of the newly described Pliocene species continues to be a subject of debate. Much of it boils down to the thorny issue of the use of species concepts within hominin paleontology, and whether or not the variation in the hominin record matches what we presently seen amongst extant primates (particularly apes). As the hominin evolutionary tree becomes bushier and bushier, particularly in the Pliocene, researchers are increasingly having to rectify these issues with the very nature of adaptive radiation and the tempo of evolution at both a micro and macro scale (Tattersall 1986; Tattersall 2000; Potts 1998; Foley 2001; Haile-Selassie 2016).
26


Within studies of intraspecies variation, there has been little consensus on the acceptable degree of morphological variability and body size dimorphism in hominins, even early Homo, but it is particularly notable with Au. afarensis. In addition to the early debates and controversy regarding the inclusion of the Laetoli and Hadar collections within a single species (Johanson and Edey 1990), there has been differences of opinion regarding whether or not the Hadar Au. afarensis collection comprises more than one species of hominin, or is simply a high degree of morphological variation and dimorphism in a single evolving lineage (Harcourt-Smith and Aiello 2004). Multiple hominin species occupying the same site contemporaneously is well-documented within the fossil record, particularly at Olduvai Gorge and Sterkfontein, which has made taxonomic determination of post-cranial remains somewhat difficult for researchers (McHenry 1994). The difficulties in parsing out the early Homo taxonomy, with H. habilis, H. rudolfensis, and H. erectus sensu lato overlapping both temporally and spatially in Pleistocene Africa, is also a continuing source of debate.
In the case of Au. afarensis, it is generally conceded that Hadar constitutes a single lineage evolving over the course of ~400ka. If this is the case, then Au. afarensis represents a highly sexually dimorphic species, with a relatively large amount of morphological variation. The sheer amount of hominin fossil material from all stratigraphic layers of the Kada Hadar member, particularly that collected since the mid 1990s return to the site, has been argued to demonstrate the time-averaged variation in the lineage, filling in many gaps left by the original round of research (Johanson 2004; Kimbel and Delezene 2009; Ward, et al. 2012). However, there is continued disagreement about the relative level of skeletal dimorphism seen at Hadar. Some authors view the dimorphism as on par with that seen in modern H. sapiens (Reno, et al. 2003; Reno, et al. 2010), while other studies have shown an elevated level of dimorphism above modern H. sapiens and Pan (McHenry 1991; Ruff 2002), or on par, or approaching, that of modern Gorillas (Richmond and Jungers 1995;
27


Johanson 2004; Plavcan, et al. 2005; Grabowski, et al. 2015). Recent discoveries at Woranso-Mille, Ethiopia (Haile-Selassie, et al. 2010) could lend further credence to the latter theories, as the partial skeleton uncovered there has a reconstructed stature far exceeding any other known Au. afarensis individual. Outside of skeletal size dimorphism, some authors have argued that there may potentially be more than one form of locomotor repertoire displayed in the Hadar assemblage, pointing towards different species, although it is argued that there is little evidence to support these claims (Harcourt-Smith and Aiello 2004).
Plavcan, et al. (2005) and Ackerman and Smith (2007) nevertheless urge caution when approaching research questions of sexual dimorphism and specific variation, as there still exists gaps within the fossil record, and thus our understanding.
The issues explored in this chapter all come to a head when interpreting the Laetoli Site S hominin prints and the questions that have emerged from their discovery. The size differences between the Site S and Site G prints are notable, but are their morphological and gait differences between them? Would any morphological differences suggest functional differences, or simply normal variability in foot construction? Can the Au. afarensis hypodigm subsume an even greater amount of body size and morphological variability, or is their greater hominin diversity at Laetoli than previously assumed? Could the paleoenvironment of Pliocene Laetoli have supported more than one hominin species, and to what extent would they have been divided adaptively and behaviorally? If the Laetoli prints were made by either one or two hominin species, what are the evolutionary and behavioral significances for the populations at Laetoli and early hominins in general? Footprints are immensely useful, but caution needs to be taken when attempting to make further reaching interpretations based on them. However, when combined with additional fossil data and context, wider patterns can potentially be seen.
28


CHAPTER III
MATERIALS AND METHODS
Materials and Data Collection
Sample
The thirteen hominin footprints from Site S are located within three separate trenches, named TP2, M9, and L8 by the initial impact assessment team, as described in Masao, et al. (2016). The three trenches were re-opened under the direction of Charles Musiba during the summer 2016 field season, as part of the UNESCO-recommended regular monitoring at the site, in order to assess the trackway and capture the photographic data utilized in this analysis. The footprints themselves are of highly variable preservation conditions. Several prints are incomplete, partially faulted or distorted, or filled with dense, difficult to extract sediment, similar to that seen in several of the prints in the G Trackway (White and Suwa 1987). Nevertheless, some of the prints are in better preserved, more informative conditions. The comparative data from the Site G trackways were taken from the first generation cast of the southern portion of the trail housed in the Laetoli visitors center at Locality 8, also during the summer 2016 field season. Later hominin and modern human comparative data was culled from published data from numerous footprint related studies, including measurements from lleret, Kenya (Dingwall, et al. 2013), Peru (Tuttle, et al. 1990), and worldwide (Tuttle, et al. 1987). A substantial comparative data set of modern unshod human footprint metrics from the Hadzabe of Tanzania, collected by Dr. Musiba used in the Musiba, et al. (1997) study, was given for use in this research. The use of data from habitually unshod or minimally shod modern humans provides a much more directly comparable data set to use with early hominin footprints, as it has been demonstrated that habitual footwear use results in different foot morphology, particularly in foot breadth and relative hallucial gap (Tuttle 1990). Metrics included in this study were footprint length, heel
29


width, ball width, foot angle, step length, stride length, stride width, and stature from 52 adult and juvenile individuals of both sexes.
Descriptions of Site S Hominin Footprints
Until the full extent of the Site S trackway and total number of prints in sequence can be determined, maintaining the naming structure of the individual prints as laid out by Masao, et al. (2016) is best for the sake of consistency. The prints are designated by the individual track-maker (of which there are arguably two) and their step sequence, and qualified by the trench from which they are located.
TP2 S1-1: This is the best preserved print of the three from the TP2 trench, and it has minimal fracturing through the print itself. The posterior portion of this right footprint is cut through by a calcite-filled vein running through the heel somewhat mediolaterally, as is the medial anterior portion, although these veins appear to have caused little distortion to the morphology of the print itself. The print is characterized by a markedly anteriorly placed hallucial imprint, as well as a noticeable posterior heel drag mark. There appears to be a small bovid print overlaying the print at the anterior portion of the drag mark/posterior portion of the hominin print.
TP2 S1-2: Roots and an east-west fault have affected the preservation of this left footprint, and the anterior portion is shifted downwards and morphologically distorted. Only some basic measurements could be taken as the state of preservation would prevent any detailed morphological or locomotor analyses.
TP2 S2-1: The smallest of the Site S footprints found so far, it is comparable in size, but generally larger than some of the G-2/3 footprints, and is clearly from another individual. It is noticeably distorted, as the anterior portion is remarkably broad. This could be from the individual slipping slightly during gait, or taphonomic factors could be the culprit, as there is a significant amount of roots crossing through the print itself. Masao, et al. (2016) have
30


listed this as a right print, although the preservation status has made this somewhat indeterminate. The lack of a corresponding second footprint along the line of progression lends credence to their identification. However, the uplifting in the trench directly above the print should not be discounted.
M9 S1-1: This left footprint was still largely filled with dense, compacted sediment at the time of data collection, and the anterior portion is crisscrossed with two veins filled with calcite. Basic measurements could still be taken, however, and once the adhering matrix eventually is removed, more detailed analyses can likely be done.
M9 S1-2: This right foot print is remarkably well-preserved, despite the calcite-filled fractures on the anterior portion. The morphology and shape of the print is well defined in the tuff layer, allowing for some kinesiological inferences. The hallucial impression is noticeably anteriorly placed, but not as much as in TP2 S1-1, and does not appear to be divergent, although the unfortunate placement of calcite vein may shroud this.
M9 S1-3: Similar to the previous print in the sequence, this left footprint is also well preserved, although diagonally cut across by a calcite-filled vein that unfortunately terminates and splits through the hallucal imprint. Despite this, the hallux does not seem to project as much as in M9 S1-2.
M9 S1-4: As in M9 S1-1, this print is still filled with the dense compacted matrix (if not more so). Although if the previous two prints in the sequence are any indication, once this matrix is carefully removed, the print will display the same degree of preservation. There are still two calcite-filled veins cutting through the print, however, one diagonally from the what is likely the lateral ball of the foot, through the edge of ray IV or V.
L8 S1-1: While some of the L8 prints are decently preserved, they are generally more affected by taphonomic factors than those in the other trenches, likely a result of the footprint tuff's proximity to the surface in the L8 trench. L8 S1-1 is a right footprint, heavily eroded and bearing vein distortion. Few kinesiological inferences can be made, although
31


some weight-distribution characteristics and basic length and width measurements can be taken.
L8 S1-2: This left footprint is crisscrossed with veins, and has some posterior erosion damage, but the shape and morphology of the print is decently preserved. There is some degree of medial slippage/displacement the hallux, but minimal to no noticeable divergence.
L8 S1-3: This right footprint is the best preserved in this sequence. Although it has some vein damage, particularly on the lateral anterior side, the morphology of the interior of the print is much more intact. The medial ball imprint is particularly prominent, and as with the other prints in this sequence, the hallux shows no divergence, and little protrusion from the rest of the anterior print.
L8 S1-4: This left footprint is much more fractured than the previous two, although the shape of the print has remained intact. There is some semblance of toe imprints remaining in the anterior portion, and there is a marked degree of hallucial slippage medially. It is more prominently than in S1-2, although taphonomic factors may have exacerbated its appearance.
L8 S1-6: Due to extensive faulting and tree root damage, the northern portion of the trench has a precipitous drop in footprint preservation. S1-5 in this sequence is completely obliterated, while S1-6 is only preserved by the posterior heel portion, which is barely distinguishable amongst the cracked ash fall tuff layer.
L8 S1-7: This right print has retained most of its shape, although it is heavily eroded and root damaged, particularly in the anterior portion. Basic measurements could be taken, although there was some estimation with the anterior section. The footprint tuff itself is heavily fractured in this section, and the sequence presumably ends here.
32


Figure 3.1: Footprints from TP2 Trench. Clockwise from top-left: TP2 S1-1, S1-2, and S2-1.
33


Figure 3.2: M9 Trench Footprints. Clockwise from top-left: M9 S1-1, S1-2, S1-4, S1-3.
34


Figure 3.3: L8 Trench Footprints. Clockwise from top-right: L8 S1-2.S1-3, S1-4.
35


Figure 3.4: L8 Trench Footprints. Clockwise from top-right: L8 S1-1, S1-6, S1-7.
36


Photogrammetric Methods
Use of photogrammetry in prehistoric footprint studies. The use of digital photogrammetry in paleontological and archaeological studies has greatly increased in prominence in recent years. Advances in capture and processing technology and software have substantially increased the capabilities of researchers utilizing photogrammetric methodologies, as well as reducing the overall cost and difficulty of its use. These advances have been a boon especially for studies of prehistoric footprints, where understanding both the extent and depth profiles of trackways is particularly crucial for a variety of research questions (Matthews 2008; Matthews, et al. 2016). The use of manual photogrammetric methods to create contour maps of footprints was pioneered by Leakey and Harris (1987) on the Site G hominin trackway, and demonstrated the importance of photogrammetry in providing crucial visual datasets for trackway studies (Bennet and Morse 2014). Digital capture methods, both laser scanning and digital photogrammetry, allow for greater accuracy in reconstruction in three dimensions of footprints, as well as the ability to conduct analyses that would otherwise be difficult or impossible to do on the hard prints themselves, and visualize qualitative characteristics that are harder to discern in the sediments. The ability to create numerous high resolution three-dimensional and orthographic models allows researchers to better visualize larger scale movement patterns of prehistoric fauna, including hominins, that would otherwise be time-consuming and difficult to plot accurately. While there are slight differences in the results between optical laser scanning and close-range digital photogrammetry, the latter has the added benefits of ease of transportation and use and cost-effectiveness, as it can be done with a relatively simple setup and a decent high-resolution digital camera (Matthews 2008; Bennet, et al. 2013, Bennet and Morse 2014; Matthews, et al. 2016). This is particularly useful for footprint sites, which are often located in remote areas with harsh conditions that would increase the difficult of using bulky, sensitive technology such as laser scanners.
37


In addition to the utilitarian and direct data collection benefits of using digital photogrammetry to analyze footprint trackways, its use is crucial for the preservation and conservation of prehistoric footprint trackways like those at Laetoli. Fossilized footprint trackways, having been imprinted upon an ever-changing landscape, are particularly vulnerable to the forces of geology and the environment, including erosion, perturbation, and uplift. At Laetoli, the situation for the innumerable animal trackways is noticeably dire. Even with the conservation steps taken on the Site G hominin trail, it has nevertheless been affected by all of the above forces, and the integrity of the prints has lessened since their original discovery. While strides were taken to protect the Site G hominin trails, the associated animal trails have been continuously threatened from pastoralist herding, erosion, and the intrusion of vegetation into the footprint tuff (Musiba, et al. 2008). Like other fossilized footprint trails, the Laetoli trackways are an immensely important data set that provide insight into the faunal community and the paleoecological makeup of an area (Lockley 1998; Musiba, et al. 2008). Since there is no way to prevent the natural forces affecting the trackways, photogrammetry is uniquely suited to help preserve the footprint record in some way for future researchers. Close range photogrammetry has been used not only to document paleontological and cultural sites for posterity and preserve the data digitally, but can also be used to document the effects of natural and anthropomorphic forces on the sites (Bates, et al. 2008; Matthews 2008; Bennet, et al. 2013; Matthews, et al. 2016). Both the Site S and Site G hominin trails are closely associated with the trackways of other animals, and photogram metric methods can preserve this relationship, even after erosion and trampling have taken their toll.
Photo capture. The nature of photogrammetric data collection necessitates photo capture within well-lit settings and from multiple angles, in order to sufficiently capture the details and depth of the subject. In the case of footprint trackways, the subject can be
38


treated as a miniature landscape. The affects of lighting and shadows on depth/height reconstruction, as well as accuracy need to be accounted for, in order capture the best possible photograph. The photos of G Trackway cast were captured by myself and Dr. Musiba with a Pentax K10D camera. The camera was mounted upon a roughly two-meter telescopic boom which allowed for consistent photo-capture height and horizontal positioning by the handler. The camera was also equipped with a remote shutter with which a second person operated the photo capturing, allowing the camera handler to maintain their positioning and control photo overlap consistency. After reviewing and discarding photos with poor lighting and resolution, well over 200 photos were captured and utilized for the G trackway cast. The Site S trackway photos were similarly captured on the telescopically mounted Pentax K10D, although I was unfortunately not present for the reopening of the trenches and subsequent photographing, and other members of the team were responsible for the data capturing. Between the three trenches, roughly 300 images were captured for photogrammetric purposes, with roughly half of those dedicated to the larger L8 trench. In all cases, meter-sticks and small scales were placed in order to provide calibration for the metric data in the resulting photogrammetric analysis.
Table 3.1 Photogrammetric Capture Data
Track way Capture Heights (m) Avg. Focal Length (mm) Image Resolution (pix) Pixel Size (mm2) # of points in cloud Avg. Root Mean Square Reprojection Error (pix)
SiteG Cast 1.0/2.0 33/18 3872x2592 0.00619046 21.6mil 1.24
L8 1.0/2.0 55/18 3872x2592 0.00619046 22.9mil 2.70
M9 1.0m 55 3872x2592 0.00622821 16.9mil 0.412
TP2 1.0m 24 3872x2592 0.00619046 8.5mil 0.201
39


Photo processing: Two separate programs, Agisoft Photoscan (v. 1.2.4) and
Photomodeler Scanner, were utilized to create accurate three-dimensional photogrammetric models of the trackways. These photogrammetry programs function by recognizing duplicate features and points across numerous overlapping photographs, which allow for the creation of an extractable XYZ "point cloud" with which to build a three-dimensional model from. Both programs use similar processing procedures, but function slightly differently from one another, which allowed for a control-check on the different models created by the programs. The functionality and capabilities of the programs also influenced what data was extracted from each of the models. These programs allow for a large degree of automation in photogrammetric procedure, as well as manual tweaking and correction of the resulting point clouds and 3-D models. Since no control-point calibration was conducted prior to photo capturing, the combination of automated algorithms and manual correction of the models was incredibly useful. The procedure for processing photos in these programs was as follows:
1. Orientation of inputted photos and estimation of camera positions
2. Alignment of photos
3. Construction of sparse point cloud
4. Manual editing of sparse point cloud to remove any incorrect/outlier points
5. Manual marking of points on the photos and creation of known distance scales
6. Recalibration and alignment of photos based on created scales
7. Creation of dense point cloud surface
8. Further manual editing of dense point cloud surface for accuracy and unnecessary point reduction
9. Triangulation and creation of three-dimensional mesh surface.
40


Figure 3.5: Digital elevation model of the southern portion of Site G Trackway cast
41


Figure 3.6: M9 Trench photogrammetric contour plot. Bar across the trench diagonally is a
two-meter stick.
42


Figure 3.7: TP2 Trench photogrammetric contour plot with digital elevation color map. The tan rectangle denotes the scale used in the model.
43


Figure 3.8: L8 Trench photogrammetric orthophoto showing direct textures from photos. Note the abundance of faunal tracks also present.
44


Numerous iterations and edits of the point-cloud meshes were done to achieve the
most accurate and usable three-dimensional models of the trackways. Longer processing times and frequent technical difficulties arose from Photomodeler, which could potentially be the result of available computing power, or the nature of the data set itself. Additionally, Agisoft Photoscan consistently produced more accurate models that required less manual editing and re-processing. As such, practically all data utilized in this study was extracted from the processing conducted in Agisoft Photoscan, with the exception being the Site G trackway cast. Photomodeler largely functioned as an accuracy check, as its recognition of problematic images, typically from resolution or poor overlap, was often more augmented than Agisoft Photoscan. Upon the recognition of missing points in the cloud, or poorly aligned pictures in Photomodeler, they could be corrected quickly in Photoscan.
Post-processing and measurement data extraction: Following the creation of the dense point cloud and mesh within Agisoft Photoscan, two additional steps were taken to allow for metric data extraction. The XYZ point cloud coordinates for each entire trackway were exported into ArcMap (v. 10.5), where a color-coded elevation maps and contour interval overlays were created. This allowed for heighted visibility of both the hominin and associated animal trackways, and the subtleties contained within the footprint tuff, as well allow visualization of relative depth and weight transfer within the hominin prints. While ArcMap allows for a great deal of resolution, it is better suited for broad, overarching visualizations and measurements of the trackways. Individual contour elevation maps were created for each hominin footprint by extracting their limited XYZ coordinates and uploading them into FootProcessor, an open-source program designed by the Bournemouth University team who worked on the Happisburgh footprint trackway, specifically for visualizing and measuring footprint data. In addition to individual contour maps, landmarks on the footprints themselves could be more accurately located and measured, and the XYZ data for the landmarks were extracted.
45


Comparative Analyses
Footprint and Trackway measurements
Standard measurements of the hominin footprints were taken following the methods used by Bennet and Morse (2014) (Fig.3.9). For consistency, all measurements were taken from the rim of the footprint. Footprint length was defined as the length from the pterion to the maximum extent the second digit. Hallucial length can vary significantly even in modern human populations, as well as in an individuals' footprints themselves, based on substrate mechanics, speed, gait, etc, thus using the second digit as the terminus of the forefoot can provide more consistency in measurement. Maximum heel width and ball width were both taken from the midline (typically the greatest extent of their respective imprints, perpendicular to the longitudinal axis defined the foot print length). Instep width was calculated as the distance between the deepest point of sole pressure and the external edge of the midfoot at its minimal breadth. Step length and stride length were measured as the distance from heel to heel, and stride width was calculated using the same landmarks.
Angle of gait was measured in ArcMap by marking the line of progression and the footprint length, utilizing right triangle plane geometry to calculate the angle. Foot index is calculated as the maximum ball width divided by the footprint length.
46


Figure 3.9: Measurement schemes utilized. Images from Bennet and Morse (2014).
Given the overlapping nature of the G-2/3 footprints, only a portion of the G-3 prints were defined enough to allow for more accurate measurements to be conducted, other than stride/step lengths. Even then, it is noted that the deformation caused by the overlapping nature may in fact bias certain measurements, such as heel and ball width. Given the overall similarity, size notwithstanding, of the G-1 and G-2/3 prints, this bias would arguably not affect the comparative analyses significantly. Additionally, even with photogrammetric imaging, it is difficult to parse out with any certainty accurate measurements of the G-2 individual, so only measurements of G-3 were utilized.
Metric Comparison with G Trackway and Modern Humans
In addition to a visual qualitative comparison between the photogrammetric plots of the Site S and Site G footprints, the combined metric data was transferred into Past (v. 3.0),
47


a statistical program designed for use with paleontological data. The footprint length and estimated stature of the S-1 and S-2 individuals were plotted alongside published data on modern humans, the Site G tracks and the lleret trackways (Tuttle 1987; Tuttle 1990;
Musiba, et al. 1997, Dingwall, et al. 2013; Bennet, et al. 2016). Two multivariate analyses were conducted utilizing the Site S, Site G, and Hadza gait study data. A principal components analysis (PCA) was conducted on the sample to determine the variables and combinations of variables most responsible for the variation in the sample. Additionally a multivariate analysis of variance test (MANOVA) was conducted on the metric data including the PCA variables to determine how statistically significant any variation seen between the groups were. A discriminant function analysis (DFA) was then done to test the placement of the S-1 individual against the G track-makers based on the metric and principal component variables. S-1, G-1/G-3, and the modern humans were placed in their own separate groups. If S-1 displays closer affinities to either the G-1/3 or the modern human data, it could provide insight into differences in foot morphology, gait, and function.
Procrustes Landmark/Shape Analysis
While the differential preservation of the S-1 footprints precludes certain kinesiological metrics, particularly in three-dimensions, a comparison of the overall shape and location of crucial landmarks between the Site S and Site G footprints is capable of being conducted. Given that foot function influences the morphology of created footprints, comparing the morphological differences between the S-1 prints and G-1/G--3 can potentially be functionally illuminating. Following similar procedures laid out in Berge, et al. (2006), 13 homologous landmark points (Table 3.2) were placed on the footprint contour maps within FootProcessor, which in addition to providing inter-landmark distances, also allows for extraction of XYZ coordinates for the placed landmarks. Landmarks were chosen based on their importance in footprint morphological features, as well as their visibility within
48


the S-1 sample, many of which are less defined than within the Site G footprints.
The six best preserved prints of the thirteen Site S tracks were chosen, TP2 S1-1,
M9 S1-2, M9 S1-3, L8 S1-2, L8 S1-3, and L8 S1-4 (Fig 3.7). The other tracks were excluded for poor preservation (TP2 S1-2, L8 S1-1, L8 S1-6, L8 S1-7), or their inability to be examined fully due to adhering matrix (M9 S1-1, M9 S1-4). TP2 S2-1 had to be excluded from the analysis because of its marked deformation, preventing accurate placement of several key landmarks. Ten prints in total were chosen for the G Trackway sample, five from each set of prints, G1-29, G1-28, G1-27, G1-33, G1-36, G3-18, G3-24, G3-26, G3-27, and G3-28 (Fig. 3.8)
Table 3.2: Landmark placement
Landmark Location
L1 Posterior Heel
Edge
L2 Anterior Toe II
L3 Medial Heel Edge
L4 Lateral Heel Edge
L5 Medial Ball Edge
L6 Lateral Ball Edge
L7 Anterior Hallux
L8 Mid sole
L9 Lateral sole
L10 Lateral edge of toes
L11 Heel center of
pressure
L12 Mt I impression
L13 Medial Sole
49


In order for the prints to be compared without the bias of relative size influencing the analysis, the landmark coordinates were converted in Past into Procrustes coordinates. The process of Procrustes fitting translates, rotates, and scales the shape coordinates, superimposing them, and thus making them directly comparable. Three different tests were conducted with the Procrustes fitted coordinates. The first was a point kernel density procedure to visualize the relative placement of the landmarks for all prints in the sample, allowing for direct visual comparison of the mean shape of each set (S-1, G-1, G-3, G-combined). The procedure creates a heat map showing the most frequent areas of landmark placement within and between samples, creating a basic mean outline shape for the footprints. The second was a two-dimensional relative warp principal component analysis and thin-plate spline deformation, which examines the variance seen between the sample as a function of mean shape. The relative deformations between the samples point towards the landmarks of the foot most responsible for the differences between the samples. Given the small nature of the sample, it is likely the differences between the Site S and Site G footprints are subtle. Following the creation of the new variables from the PCA, a MANOVA test was conducted to determine if the variability between the S-1 and G-mean footprints was statistically significant.
50


Figure 3.10: S-1 Prints Utilized in the 2-D Landmark Analysis. Clockwise from top-right: L8 S1-2, L8 S1-3, L8 S1-4, TP2 S-1, M9 S1-2, M9 S1-3.
51


Figure 3.11: Sample of Site G Tracks used in 2-D Landmark Analysis. Clockwise from top-right: G1-28, G1-29, G1-33, G2/3-18, G2/3-24, G2/3-27.
52


CHAPTER IV
RESULTS
Qualitative and metric comparison with other hominins and modern humans
It is obvious upon first glance that the S-1 track maker is much bigger than the individuals responsible for the Site G trackway. The S-2 individual's print is also noticeably larger than G-1 or G-3, although possibly close to the highest estimates for the prints of G-2. The S-1 prints average 251mm in length (Table 4.1), ranging roughly 10mm across the twelve measurable prints in the three trenches, compared to a 180mm average foot length for G-1 and 209mm average for G-3. This situates them with the upper reaches of adult Hadza males, while the Site G individuals are roughly equal to the Hadza juveniles (Fig.
4.1). Their size also translates into total breadth, with an average maximum heel width of 72.4mm and an average maximum ball width of 97.2mm. The footprints are comparable in size to the larger individuals from the lleret, Kenya trackways, and can be situated comfortably within sizes of modern human male averages across the globe, including that of the habitually unshod Hadza (Fig 4.3).
Table 4.1: Mean Metric Results
Sample Foot length (mm) Max Ball Width (mm) Max Heel Width (mm) Step Length (cm) Stride Length (cm) Stride Width (cm) Foot Angle Foot Index Stature (cm)
S-1 251.8 97.2 72.4 57.9 115.2 75.3 6 38.7 167.3 (est.)
S-2 225 N/A 71 N/A N/A N/A N/A N/A 150.6 (est.)
G-1 180 78.9 65.8 41.6 82.9 68 18.6 43.8 133.6 (est.)
G-3 209 85.5 75.9 43.3 87.6 106 -4.3 41.5 148 (est.)
Hadza (AD) 237 88.8 58.8 66.4 131.3 61.3 6.2 37.3 155.2
Hadza (JV) 212.2 78.8 50.9 59.2 118.9 61.9 6.6 37 135.1
53


Figure 4.1: Box-plots comparing footprint length of the Hadza with the Laetoli individuals.
The left plot shows the distribution based on sex, while the right shows it based on age.
In both average step length (57.9cm) and average stride length (115.2cm) the Site S individual also exceeds those of G-1 or G-3, and is comparable to the low end of variation for adult Hadza, but in the upper reaches of variation for juvenile Hadza. The Site G individuals, however, are at the very bottom reaches of Hadzabe variation, below the averages of even the juvenile Hadza. (Fig 4.2). The average stride width of 75.3mm exceeded that of the G-1 individual (68mm) and the Hadza (61,9mm), but is less than that seen in the G-3 prints (106mm). However, all of the early hominin stride widths fit within the range of variation seen in the Hadza individuals who participated in the study, which varied widely from individual to individual. With an average 6 angle of gait, the S-1 individual walked more out-toed than G-3 (average of-4.3 in-toeing), but significantly less out-toed than the G-1 individual (average 18.6), and is comparable to average angle of gait seen in the Hadza sample, who seem to generally walk with some degree of out-toeing.
54


Figure 4.2: Box-plots comparing step-length of the Hadza with the Laetoli individuals. The left plot shows the distribution based on sex, while the right shows it based on age.
The Site S footprints are remarkably similar in their visible qualitative functional qualities to the Site G prints, but with some notable differences. They possess a deep, ovalshaped heel strike impression followed by movement along the lateral side of the sole, with transfer of pressure medially along the metatarsals. There are variable impression depths alongside of the medial sole, but clear evidence of a medial longitudinal arch with the deepest impression of the midfoot occurring on the lateral edge. As seen in many of the G-1 prints, the deepest part of the metatarsal impressions in several of the prints occurs on the lateral side, and the lateral toe imprints are noticeably shallow, although more prominent in S-1. Unlike the G prints, the hallucial imprint is less defined, with the exception of TP2 S1-1 and M9 S1-2, although this may be the result of taphonomic factors, particularly in the partially fractured L8 trench prints. However, the hallucial imprint is noticeably deeper. Compared to the G-1 prints, the hallux does not appear to diverge as much from the second
55


digit, TP2 S1-1 and possibly M9 S1-3 excluded.
When the stature estimates and foot length of the Site S individuals are plotted alongside those of the Site G prints, lleret footprints, and global modern human data from Tuttle (1987) (Fig. 4.3), the S-1 individual is situated well within the distribution of modern human males and the lleret track-makers, while the S-2 individual is in the upper-mid range for modern human females, and near the uppermost estimate for G-3 track-maker (Bennet, et al. 2016). Even at the lowest stature and size estimate for the S-1 individual, it is clear that they were likely significantly larger than either of the Site G track makers.
2.240 2.224 2.208 2.192
3 -
2.176
3
2.160 2.144 2.128 2.112
2.22 2.25 2.28 2.31 2.34 2.37 2.40 2.43 2.46 2.49
Log Foot Length
Figure 4.3: Stature and foot length plot comparing hominin individuals with global data of
modern humans.
The principal components analysis (PCA) showed that two principal components were largely responsible for the variation in the sample, the remainder evenly distributed through an additional seven principal components. The variables utilized in the PCA were stature, footprint length, heel width, ball width, step/stride length, and foot index. Foot angle
w Xileret FUT1B XileretA2-13
ITIA^tere^i*
0 >)ler lem Human Males
lleret A2-12
G-
+Hadza1WsaMales (Bennet 2016 esL)
O S-2
+HadzaFemal^
G-3 (Adult Hadza est.H3-3 (Adult Mach, est.) M
Modem Human Females
CP
o o
0
o
3-1 (Adult Mach, est.)
3-1 (Adult Hadza est.)
3-1 (Bennet 2016 esL)
+Hadza Juveniles
56


and stride width were excluded from the final analysis, as their within-group and individual variability was creating noticeable skewing of the results. PC1, with an Eigenvalue of 656.2 explained 70.3% of the variance, while PC 2, with an Eigenvalue of 178.3, explained 19.09% of the variance. Based on the loadings, PC1 largely represents on footprint length and gait, while PC2 largely represents the expected gait given the overall sizes in the sample. The PCA plot (Fig. 4.4) shows a clear separation between the G, and S-1 samples, while the Hadza maintain an even, but clustered distribution across the center of the plot, which is to be expected considering the mix of men, women and juveniles in the sample. The S-1 tracks show a much closer affinity to the Hadza sample than to the Site G individuals.
Figure 4.4: PCA Plot with 95% ellipses showing all footprints used in the sample.
The Discriminant Function Analysis (DFA) (Fig. 4.5) maintains this affinity, demonstrating the degree to which S-1 clusters with the Hadza sample. The first function, which corresponds with PC1, displayed an Eigenvalue of 1.673, while the second function, corresponding to PC2, displayed an Eigenvalue of .574. The Wilks Lambda, which was
57


below 1, showed the separation between the three groups along these two axes was statistically significant, but not highly discriminatory (A=.238, p=<.005, df=4). However, the relatively high canonical correlation for both functions (r=.791 and r= .604 respectively) demonstrates that there is a strong association between the discriminant scores and the groupings. The DFA largely placed the three groups within their own classifications, particularly the Site G group, which remained entirely in its original classification. The Hadzabe sample had several prints classified within the S-1 group, which, given their degree of overlap, is to be expected. Several Hadza prints were classified within G, either from juveniles or a single adult outlier.
Figure 4.5: DFA plot with 95% ellipses. Note the adult Hadza outlier's affinity with Site G.
The MANOVA of the sample based on the first two principal components confirms the statistical significance of the variance within the groups (p=<005, A=.763, F (2, 86) = 13.357). Interestingly, when the three groups are compared against each other in posthoc tests, there are different significances based on the two principal components. With PC1, both S1 and the Hadza samples are significantly different from the G sample (p=<.005), but
58


not from each other (p=.530). However in PC2, the differences between all three groups is statistically significant (p=.001).
Kernel Density and Procrustes Landmark Analysis and PCA
The kernel density plot of the 13 landmark point shows clear differences in the breadth, particularly in the midfoot and forefoot region, of the S-1 (Fig. 4.7) and Site G (Fig.4.8) footprints, even when controlling for overall size. The below density plots show the areas of greatest concentration for landmark placement, visualized as a heat map. Additionally the G-1 and G-3 prints show a greater degree of divergence between the first and second digits than in S-1, as well as a somewhat less anteriorly placed center of heel pressure. The results of the two-dimensional PCA of the relative warp between the two samples show a much more even distribution of variance than seen in the metric data PCA, with greater overlap between the two mean footprint shapes (Fig. 4.6).
Fig. 4.6 Plotting PC1 Vs. PC2
59


Fig. 4.6: S-1 Print Landmark Kernel Density Plot.
Fig. 4.7: G-1/G-3 Avg. Landmark Kernel Density Plot.
60


Figure 4.9: 2-D PCA Relative Deformations. Clockwise from top right: PC1, PC2, PC3, PC4.
PC 1 (Fig.4.9), with an Eigenvalue of .0026 explained 26.8% of the variance, largely focuses on midfoot breadth and hallucal gap, based on the relative deformation plot. PC 2 (Fig. 4.9), with an Eigenvalue of .00194, explaining 20.02% of the variance, PC3 (Fig 4.9.), with an Eigenvalue of .0013, explaining 13.6% of the variance, both seem to center around overall breadth of the foot from different landmark distances. PC4 (Fig. 4.9), with an Eigenvalue of .0011, explaining 11.8% of the variance, seems to center exclusively on instep width.
61


Figure 4.10: Overall Deformation of Mean Shapes from S-1 Avg. to G-Avg.
The overall thin-plate spline deformation plot (Fig. 4.10) confirms that hallucal gap and midfoot breadth are the primary differentiating variables in the mean shapes of the S-1 and G-1/G-3 individual. The low Eigenvalues and relatively even distribution of variance across multiple principal components confirms that the differences between the mean footprint shapes are subtle. The MANOVA on the PCA results show the differences are statistically significant, but with weakly associated variation, with the Wilks' lambda test showing a p-value of .002 (A=.336, F (4, 15)=7.426).
Summary of Results: The combined PCA and DFA tests show that there are significant morphological differences separating the S-1 and G footprints. While there are similarities in foot pressure distribution, and in stride/step lengths, in overall metrics and principal components, the S-1 individual aligns more closely with the Hadza sample than with the Site G individuals. In mean footprint shape, the S-1 individuals show a much narrower foot, particularly in the midfoot, and a less diverged hallux. These differences are significant enough to discriminate between the two samples.
62


CHAPTER V
DISCUSSION
The results of the comparative print analyses are incredibly intriguing in their implications. While the sheer size of the prints, as well as the reconstructed stature of the S-1 individuals demonstrate that this was clearly a different, noticeably larger, individual than the Site G track makers, there is also potentially differences in foot morphology and locomotion. The Site S sample, while overlapping with both some of the Hadza sample and with the larger G-3 prints (and theoretically the G-2 prints), consistently plots in between the two samples in both the PCA and DFA largely skewing closer to the Hadza sample, especially in the DFA. The PCA plot shows the three samples distributing along the axes based on foot size and expected mobility. The Hadza sample shows a fairly even linear distribution along the center of the intersection of the two axes, with the juveniles, possessing smaller feet and the expected shorter strides to the left, and the adults with larger feet and larger strides to the right of the plot. The G-1 and G-3 individuals both cluster on the left side of the plot, reflecting their smaller size and smaller strides. The G-1 and G-3 prints form their own individual clusters, based on the size differences between the individuals. Whereas based on their size, the G-1 prints can be expected to shorter strides, albeit ones shorter than a modern human of equivalent size, the G-3 prints have similar relative stride lengths, despite being made by a larger individual. The S-1 print cluster in a similar position on the PC2 axis as G-3, but plot further right on the PC1 axis, partially aligning with the Hadza sample. Their location on the upper end of the PC2 axis indicates that despite their large size, they nevertheless have smaller strides than would be expected for individuals of their size.
The discriminant function analysis plot based on the first two principal components (which accounted for 87.4% of the total variance) affirms these differences. The prints of the
63


two G-individuals cluster closer together along the left side of first axis, maintaining their distinctness from the S-1 and Hadza clusters. The S-1 cluster displays even greater affinity towards the Hadza sample, which is affirmed by the statistical analysis showing that these two samples are significantly more different from the Site G sample than they are from each other, at least within PC1. The predicted group membership results, seems to have differentially discriminated the outlier prints for their closer affinities with the further ends of each groups principal component variation.
The qualitative comparison shows similar foot function in terms of footprint formation, and the preservation of the currently excavated Site S prints would necessitate a very cautious approach to any pressure depth analysis, hence why only a 2-D landmark analysis was conducted in this study. However, the comparisons of the mean shapes of the footprints based on the 13 chosen landmarks demonstrate that there are subtle, yet significant differences in the footprint morphology between S-1 and the Site G individuals. The principal components analysis showed a fairly even distribution of variance across several principal components, although largely within the first four. The 2-D deformation of the principal components confirms the even distribution of variance across the whole footprint, but shows consistent differences in footprint breadth and hallucial gap. The primary point of breadth deformation is located in the medial midfoot, which would correspond with the location of the medial longitudinal arch. These significant differences between S-1 and the G individuals cannot solely be accounted for by overall size. The S-1 prints in mean shape are still relatively broad, but more closely resemble the more narrowed shape of modern human footprints as described in Berge, et al. (2006). The S-1 prints show additional affinities in their phalangeal positioning. The hallux is more closely aligned with the second digit, and the lateral toes decrease in size sequentially from the hallux, where as in the G prints, the lateral toe imprints are as long, and in some cases, longer than the hallucal imprint.
The differences in footprint shape point towards differences in foot morphology
64


between the individuals. While the differences in hallucal gap have more obvious physiological correlates, the mid-foot differences likely reflect arch development and construction, in which the S-1 individual possessed a higher, more human-like, medial longitudinal arch. While qualitatively the distribution of pressure in all three sets of Laetoli prints appear to be remarkably similar, without an quantitative analysis of pressure distribution, it is difficult to make detailed inferences about any kinematic differences between the S and G individuals. Nevertheless, in overall shape, the S-1 footprints reflect an arguably much more human-like construction to the foot, and it would not be outlandish to suggest that some functional differences would be expected. The results of the discriminant function analysis, which plotted the S-1 individuals away from both the G individuals and nearly within the Hadza sample, suggests that the S-1 individual differed in locomotor pattern from both groups, although the latter less so. The issue at hand then, is what these differences mean as far as taxonomic variation in the Pliocene Laetoli hominins.
Modern humans are not without great variation in foot morphology, function, and footprint shape. Hallucal gap, arch development, toe length, and foot breadth are all dependent on a variety of factors, both physiological and environmental. Habitually unshod populations consistently display broader feet and a larger hallucal gap than groups that consistently wear shoes, and while it is sometimes argued that they display lower arches, studies such as Tuttle, et al. (1990) and Musiba, et al. (1997) demonstrate this is not necessarily the case. There is also variation between unshod populations, as Musiba, et al. (1997) demonstrated with the Hadzabe, which on average have longer and narrower feet than the Machiguenga people studied by Tuttle, et al. (1990). Musiba, et al. (1997) also remark that the Hadza, for their stature, display longer strides and quicker gaits than seen in comparable populations. However, it is important to note that while there are some variations within populations, mostly age related, the mean of many metrics and shape-wise comparisons are largely similar between individuals of both sexes. In the comparative PCA,
65


Hadza men and women almost entirely overlapped with one another, and while juveniles skewed closer to the Site G sample, the latter still was largely out of the Hadza range of variation. Within the G prints, the variability between the prints and between the track-makers is fully within the expected variation seen in a population, based on the Hadza data. Despite the size difference between G-1 and G-3, they maintain the same mean footprint shape. Conceivably the G-2 print would fall within the same mean shape, were it better visible. While the S-1 individual displays some affinities with the G individuals, consideration needs to be made of the significant differences in footprint shape. While subtle but significant differences in footprint morphology are on the surface not incredibly differentiating, when contextualized with their size and the known paleontological record, a pattern can be glimpsed.
It is abundantly clear that the S-1 individual is a surprisingly large hominin for the Pliocene. No matter what method of stature reconstruction utilized, whether its a simplified height/foot length ratio (Tuttle 1987), the Hadza stature-based regression (Musiba, et al. 1997), or primitive body proportion-based (Dingwall, et al. 2013), S-1 is significantly larger than either G-1, G-2, or G-3, while S-2 is larger than G-1 or G-3, and likely G-2. The stature and presumptive size of S-1 greatly exceeds that of any known Au. afarensis individuals from Laetoli or Hadar. The largest known individual assigned to the Au. afarensis hypodigm is the partial skeleton from Woranso-Mille, KSD-VP-1/1 ("Kadanuumuu"), which is also notable for its size, described as being the size of a small modern human (Haile-Selassie, et al. 2016). Masao, et al. (2016) give an estimated stature of 158cm for "Kadanuumuu," likely based off the tibial length, but do not provide their method of stature estimation, as neither Haile-Selassie, et al. (2010) nor Lovejoy, et al. (2016) give a direct estimation of the stature of the individual. The implication that the Woranso-Mille individual's lower limbs were elongated, despite having an Au. afarensis-like upper body plan (Haile-Selassie, et al.
2016), may influence what the reconstructed height of this individual would be. Additionally,
66


while "Kadanuumuu" has been assigned to Au. afarensis based on shared morphological traits, no associated craniodental remains have published for the individual, which prevents definitive taxonomic assignment. However, even if one accepts this estimation, S-1 is still taller at the lowest end of its estimated stature. There are other important implications of the size and morphology of the S-1 prints.
Alongside "Kadanuumuu," the Site S hominin prints demonstrate that there were large bodied, bipedally-advanced hominins in the Pliocene. While previously "Kadanuumuu" could be construed as an exceptionally tall outlier, much as "Lucy's" diminutive stature has been, the S-1 prints confirm that this was not the case. Coupled with the existence of smallbodied hominins such as H. naledi, H. floresiensis, and Au. sediba in the Pleistocene, the notion of a linear trend of increasing body size and postcranial change in hominin evolution is becoming increasingly harder to defend. Grabowski, et al. (2015) and McHenry and Brown (2008) also both posit this from different morphological perspectives, which stresses the point that hominin evolution, particularly the transition from the australopithecine grade to Homo, was not a tidy evolutionary affair. One does not need to believe in a much more speciose hominin evolutionary tree to recognize that adaptive responses of the Plio-Pleistocene hominins was likely not driven by unidirectional forces.
However, this leads to another important, more controversial, implication of the Site S prints: the taxonomic designation of the track-makers. Given their contemporaneity with the G prints, and the latter's widely accepted attribution to Au. afarensis, it is not an unreasonable assumption that the Site S prints were also made by Au. afarensis, albeit a much larger one. While the Hadar Au. afarensis assemblage has been argued to display a noticeable amount of sexual dimorphism in body size (Kimbel and Delezene 2009), the differences within that sample are dwarfed by the estimated differences between S-1 and G-3. Presuming that S-1 was a male, and the Site G individuals were females and a juvenile, as Masao, et al. (2016) have interpreted, this is an enormous amount of sexual dimorphism
67


within a species. This level of dimorphism is not unheard of in primates, as both gorillas and orangutans have significant size differences between males and females. This alone would have significant evolutionary and possibly behavioral implications as far as social structure is concerned, which Masao, et al. (2016) subsume into their interpretation of the track site. The noted differences in footprint morphology, which suggest differences in foot structure, could suggest locomotor dimorphism, wherein the larger males were more advanced bipedally than female individuals, a notion entertained by some researchers of Au. afarensis (Harcourt-Smith and Aiello 2004). Considering the significant differences in the Laetoli track sets, and the lack of Au. afarensis fossils other than "Kadanuumuu" that approach the reconstructed size of S-1, it is imperative to consider alternate interpretations regarding the track-makers at Site S.
Given the emerging picture of hominin diversity in the middle Pliocene, both in terms of speciation and postcranial variation, the likelihood of another hominin species inhabiting Laetoli is not an outlandish proposition. Laetoli is already noted for its relative rarity of hominin remains, Au. afarensis included (Su and Harrison 2008), and the taphonomic and environmental causes for their rarity may mask taxonomic diversity. The animal trackways have provided crucial paleoecological evidence fleshing out the faunal community of Pliocene Laetoli, particularly for species that may not be fully represented by the fossil record (Musiba, et al. 2008), and it is logical that the same would be true for hominins as well. While the lleret footprints are commonly attributed to H. erectus, their discoverers note that H. habilis and Paranthropus boisei are also known to have existed in the area contemporaneously with the footprints, and are just as likely to have been the track-makers (Dingwall, et al. 2013). The existence of K. platyops and other australopithecines in East Africa in the middle Pliocene roughly contemporaneously with Au. afarensis points towards potential overlap in range and habitat among species. The discovery of the first fossils of P. aethiopicus outside of Ethiopia and Kenya in the Ndolanya Beds at Laetoli is also an
68


important development to be considered (Harrison 2011). While known estimated ages for P. aethiopicus, the oldest known robust australopithecine, range from ~2.3 to 2.7 Ma, the fossils from Laetoli are the oldest securely dated specimens at 2.66 Ma (Harrison 2011).
The earliest known fossil evidence does not necessarily correlate with its first appearance date, and the emergence of the robust australopithecines is still phylogenetically murky. The possibility of earlier hominins with affinities to P. aethiopicus existing at Laetoli contemporaneously with Au. afarensis is entirely plausible.
Paleoecological interpretations of Laetoli during the mid-Pliocene also support the notion of potential overlapping hominin populations. At Olduvai Gorge during the Pleistocene, H. habilis and P. boisei are known to have occupied the area contemporaneously, and are argued to have occupied different ecological niches. This appears to be based on dietary and habitat preferences, as isotopic analyses on dental enamel demonstrate that P. boisefs diet consisted of a substantially higher proportion of C4 grasses and sedges than H. habilis (Barboni 2014). Andrews and Bamford (2008) note the frequent association of P. aethiopicus at Laetoli with habitats dominated by C4 plants.
Given an accepted interpretation of a dynamic, mosaic-like paleolandscape that included the gradual expansion and periodic dominance of C4 grasses at Laetoli, it is possible that hominin adaptive responses to these factors resulted in speciation. Alternatively, the periodic disruption of woodland environments could have resulted in temporary range and habitat expansion for hominin species more adapted to more open grassland environments and diet. Tuttle et al. (1990) have disputed Au. afarensis as the Site G track-maker because of its pedal morphology, and posited the existence of another, more bipedally advanced hominin residing at Laetoli. Given the more advanced appearance of the S-1 prints, their closer statistical affinity with the Hadza than with the Site G individuals, and the difficulty in rectifying their size within the known size distribution of Au. afarensis, this notion is worth revisiting.
69


As it stands, there are three possible scenarios for the taxonomic designation of the Laetoli print makers. The first, posited by Masao, et al. (2016), is that both the Site S and Site G prints were left behind by Au. afarensis, with S-1 representing a large male individual, and S-2 and the Site G prints were likely left behind by female and/or juvenile individuals. This would vastly increase the known body size distribution of Au. afarensis, and suggest a gorilla-like sexual dimorphism within the species. The differences between the two sets of prints are either normal distribution influenced by allometry, or indicative of locomotor dimorphism between the sexes, with males possessing more derived bipedal capabilities.
S-1 could also potentially be a peculiar individual with much larger feet than expected. The second scenario is that another, larger, hominin species existed alongside Au. afarensis at Laetoli, and is responsible for the Site S prints. This species would potentially bear some resemblance to Au. afarensis in locomotor pattern, but display different pedal morphology, reflected in the more-human like shape and size of the S-1 prints. These individuals could potentially be an early paranthropine, K. platyops, or another gracile species of Australopithecus. A third, equally controversial, potential scenario, is that neither Site S nor Site G trackways were created by Au. afarensis. Based on the arguments provided by Tuttle, et al. (1990) and other researchers that the species was unlikely to have created the prints, another, more bipedally derived, hominin species was responsible for both trackways. Like the first scenario, the differences between the track morphologies are subsumed in normal variation or are a function of allometry. While all three scenarios are likely, this third scenario would be a much more radical shift in the interpretation of the trackways. While attribution of the Site S prints to Au. afarensis is a valid interpretation, based on the significant differences in size and morphology of these prints with the G trackway, it is equally parsimonious to argue for the presence of another, larger hominin species at Laetoli that could be responsible.
70


CHAPTER VI
CONCLUSION
The serendipitous discovery of the Site S prints is tremendously exciting. Despite being one of the most important sites for understanding hominin evolution and paleoecology and researched for decades, Laetoli is yielding surprises, even within a stone's throw of the Site G trackway, its most prominent contribution. Given the size of the site and the extent of the footprint tuff, additional hominin prints are no doubt waiting to be uncovered alongside the innumerable animal footprints blanketing many localities within Laetoli. The Site S prints have considerable implications for hominin evolution in the Pliocene, as well as the evolution of bipedality within the hominin clade. While the interpretations of their evolutionary, locomotor, and behavioral significance are open for debate, their presence arguably confirms the existence of large-bodied hominins in the Pliocene fossil record, much earlier than previously believed.
The research undertaken in this thesis has demonstrated several important results.
In addition to placing the foot size and reconstructed stature of the Site S-1 individual well into the range of modern human males, the principal component analysis and discriminatory function analysis conducted demonstrate the statistical placement of S-1 in between the Site G individuals and the modern human Hadza sample, with greater affinity towards the latter. The MANOVA demonstrated that the differences between these samples are statistically significant, particularly between S-1 and the Site G individuals. While qualitatively, the sets of footprints appear similar in form and function, the Procrustes landmark PCA showed significant differences between the Site S and Site G footprint shapes. S-1 displays greater hallucial abduction, phalangeal lengths that decrease laterally from the hallux, and a narrower midfoot. Overall the shape of S-1 appears more derived towards the modern human condition than G-1 or G-3. The differences are subtle, but are shown to be
71


statistically significant. While these could be subsumed into normal variation, it is also potentially indicative of differential foot construction and locomotor function.
When the results of this research are viewed cumulatively, and put in the paleoecological context of Pliocene Laetoli, the implications become much more complex. While size and differences in footprint shape alone cannot definitively demonstrate taxonomic differences, this research provides evidence that it is entirely plausible that another, larger species of hominin co-existed with Au. afarensis at Laetoli, and is responsible for the creation of the Site S prints. Given our current knowledge of the Pliocene fossil record and paleoecological drivers of variation and adaptation, an attribution of the Site S prints to Au. afarensis without considering alternative hypotheses is somewhat premature. Further analyses of the existing prints, as well as the uncovering of additional prints and recovery of further hominin fossils will hopefully elucidate the emerging picture of hominin locomotor diversity at Laetoli.
72


REFERENCES
Aiello, L. and Dean. C. 1990. Human Evolutionary Anatomy. Academic Press Limited.
Ackermann, R.R., Smith, R.J. 2007. The Macroevolution of our Ancient Lineage: What We Know (or Think We Know) about Early Hominin Diversity. Evolutionary Biology. 34: 72-85.
Andrews, P. and Bamford, M. 2006. Past and present vegetation ecology of Laetoli. Journal of Human Evolution. 54, 78-98.
Barboni, D. 2014. Vegetation of Northern Tanzania during the Plio-Pleistocene: A synthesis of the paleobotanical evidences from Laetoli, Olduvai, and Peninj hominin sites. Quaternary International 322: 264-276.
Bates, K. T., Rarity, F., Manning, P. L., Hodgetts, D., Bernat, V., Oms, O., Galobart, A.,
Gawthrope, R. L. 2008. High-resolution LiDAR and photogrammetric survey of the Fumanya dinosaur tracksites (Catalonia): implications for the conservation and interpretation of geological heritage sites. Journal of the Geological Society, London 165: 115-127.
Behrensmeyer, A.K. 2006. Climate Change and Human Evolution. Science. 311, 476-478
Behrensmeyer, A.K and Reed, K.E., 2013. Reconstructing the Habitats of Australopithecus: Paleoenvironments, Site Taphonomy, and Faunas. In: Reed, K.E., Fleagle, J.G., Leakey, R.E., editors. The Paleobiology of Australopithecus. Vertebrate Paleobiology and Paleoanthropology. Springer Netherlands. 41-60.
Bennet, M. R., Falkingham, P., Morse, S.A., Bates, K., Crompton, R. H. 2013. Preserving the Impossible: Conservation of Soft-Sediment Hominin Footprint Sites and Strategies for Three-Dimensional Digital Data Capture. PLoS One 8 (4): e60755.
Bennet, M. R. and Morse, S. A. 2014. Human Footprints: Fossilised Locomotion? Springer International Publishing.
Bennet, M. R., Reynolds, S. C., Morse, S. A., Budka, M. 2016. Laetoli's lost tracks: 3D
generated mean shape and missing footprints. Nature Scientific Reports 6: e21916.
Bennet, M. R., Reynolds, S. C., Morse, S. A., Budka, M. 2016. Footprints and human evolution: Homeostasis in foot function? Paleogeography, Paleoclimatology, Paleoecology 461: 214-223.
Berge, C., Penin, X., Pelle, E. 2006. New interpretation of Laetoli footprints using an
experimental approach and Procrustes analysis: Preliminary results. C.R. Palevol 5: 561-569.
73


Bobe, R., Behrensmeyer, A.K., Chapman, R.E., 2002. Faunal change, environmental
variability and late Pliocene hominin evolution. Journal of Human Evolution 42: 475-497.
Brunet, M., Beauvilain, A., Coppens, Y., Heintz, E., Moutaye, A.H.E., Pilbeam, D.1996.
Australopithecus bahrelghazali, une nouvelle espece dHominide ancien de la region de Koro Toro (Tchad). Comptes rendus de IAcademie des sciences. Serie 2. Sciences de la terre et des planetes. 322, 907-913.
Brunet, M., Guy, F., Pilbeam, D., Mackaye, H. T., Likius, A., Ahounta, D., Beauvilain, A., et al. 2002. A new hominin from the Upper Miocene of Chad, Central Africa. Nature 418:145-151.
Clarke, R., 2013. Australopithecus from Sterkfontein Caves, South Africa. In: Reed, K.E., Fleagle, J.G., Leakey, R.E., editors. The Paleobiology of Australopithecus.
Vertebrate Paleobiology and Paleoanthropology. Springer Netherlands. 105-123.
Crompton, R. H., Pataky, T. C., Savage, R., D'Aout, K. D., Bennet, M. R., Day, M. H., et al. 2012. Human-like external function of the foot, and fully upright gait, confirmed in the 3.66 million year old Laetoli hominin footprints by topographic statistics, experimental footprint formation and computer simulation. Journal of the Royal Society Interface 9: 707-719.
DeSilva, J. M. 2010. Revisiting the "Midtarsal Break." American Journal of Physical Anthropology 141: 245-258.
DeSilva, J. M., Bonne-Aannee, R., Swanson, Z., Gill, C. M., Sobel, M., Uy, J., and Gill, S. V. 2015. Midtarsal Break Variation in Modern Humans: Functional Causes, Skeletal Correlates, and Paleontological Implications. American Journal of Physical Anthropology 156: 543-552.
Dingwall, H. L., Hatala, K. G., Wunderlich, R. E., Richmond, B.G. Hominin stature, body
mass, and walking speed estimates based on 1.5 million-year-old fossil footprints at lleret, Kenya. Journal of Human Evolution 64: 556-568.
Ditchfield, P. and Harrison, T. 2011. Sedimentology Lithostratigraphy and Depositional History of the Laetoli Area. In: Paleontology and Geology of Laetoli: Human Evolution in Context. Harrison, T., Editor. Springer Science+Business Media.
Dormnguez-Rodrigo, M. 2014. Is the Savanna Hypothesis a Dead Concept for Explaining the Emergence of the Earliest Hominins? Current Anthropology. 55, 59-81.
Foley, R. 2002. Adaptive Radiations and Dispersals in Hominin Evolutionary Ecology. Evolutionary Anthropology Supplement 1: 32-37.
74


Foley, R. and Gamble, C., 2009. The Ecology of Social Transitions in Human Evolution. Philosophical Transactions: Biological Sciences. 364: 1533, 3267-3279.
Grabowski, M. Hatala, K. G., Jungers, W. L., Richmond, B. G. 2015. Body mass estimates of hominin fossils and the evolution of human body size. Journal of Human Evolution 85: 75-93.
Haile-Selassie, Y., Saylor, B.Z., Deino, A., Levin, N.E., Alene, M., Latimer, B.M., 2012. A
new hominin foot from Ethiopia shows multiple Pliocene bipedal adaptations. Nature. 483, 565+.
Haile-Selassie, Y., Gibert, L., Melilo, S. M., Ryan, T. M., Alene, M., Deino, A., Levin, N. E., Scott, G., Saylor, B. Z. 2015. New species from Ethiopia further expands Middle Pliocene hominin diversity. Nature 521: 483-488.
Haile-Selassie Y. 2016. Introduction to KSD-VP-1/1: The Earliest Adult Partial Skeleton of Australopithecus afarensis. In: Haile-Selassie Y., Su D.F., editors. The Postcranial Anatomy of Australopithecus afarensis. Springer Netherlands. 1-12.
Haile-Selassie Y., Melillo S.M., Su D.F. 2016. The Pliocene hominin diversity conundrum:
Do more fossils mean less clarity? PNAS 113: 6364-6371.
Hatala, K. G., Demes, B., Richmond, B. G. 2016. Laetoli footprints reveal bipedal gait biomechanics different from those of modern humans and chimpanzees.
Proceedings of the Royal Society B 283: 20160235.
Harcourt-Smith, W.E.H. and Aiello, L.C., 2004. Fossils, feet and the evolution of human bipedal locomotion. Journal of Anatomy. 204, 403-416.
Harrison, T. 2011. Laetoli Revisited: Renewed Paleontological and Geological Investigations at Localities on the Eyasi Plateau in Northern Tanzania. In: Paleontology and Geology of Laetoli: Human Evolution in Context. Volume 1: Geology,
Geochronology, Paleoecology and Paleoenvironment. Harrison, T., Editor. Springer Science+Business Media. 1-15.
Harrison, T. 2011. Hominins from the Upper Laetolil and Upper Ndolanya Beds, Laetoli. In: Paleontology and Geology of Laetoli: Human Evolution in Context, Volume 2: Fossil Hominins and Associated Fauna. Harrison, T., Editor. Springer Science+Business Media. 141-188.
Hay, R. L. 1987. Geology of the Laetoli Area. In: Laetoli: A Pliocene Site in Northern Tanzania. Leakey, M. D. and Harris, J. M., (Editors). Clarendon Press Oxford.
Hunt, K. D. 1994. The evolution of human bipedality: ecology and functional morphology. Journal of Human Evolution 25: 183-202.
75


Johanson, D. and Edey, M.A., 1990. Lucy: The Beginnings of Humankind. Simon and Schuster.
Johanson, D. C. 2004. Lucy, Thirty Years Later: An Expanded View of Australopithecus afarensis. Journal of Anthropological Research 60 (4): 465-486.
Johanson, D. C., Lovejoy, C. O., Kimbel, W. H., White, T. D., Ward, S. C., Bush, M. E., Latimer, B. M. Coppens, Y., 1982. Morphology of the Pliocene Partial Hominid Skeleton (AL 288-1) from the Hadar Formation, Ethiopia. American Journal of Physical Anthropology. 57, 403-451.
Kim, J. Y., Kim, K. S., Lockley, M. G., Matthews, N. 2008. Hominid Ichnotaxonomy: An Exploration of a Neglected Discipline. Ichnos 15: 26-139.
Kimbel, W.H., Lockwood, C.A., Ward, C.V., Leakey, M.G., Rak, Y., Johanson, D.C., 2006. Was Australopithecus anamensis ancestral to A. afarensis? A case of anagenesis in the hominin fossil record. Journal of Human Evolution. 51, 134-152.
Kimbel, W. H., and Delezene, L. K. 2009. "Lucy" Redux: A Review of Research on Australopithecus afarensis. Yearbook of Physical Anthropology 52: 2-48.
Kingston, J. D. and Harrison, T. 2007. Isotopic dietary reconstructions of Pliocene
herbivores at Laetoli: Implications for early hominin paleoecology. Paleogeography, Paleoclimatology, Paleoecology 243: 272-306.
Klenerman, L., and Wood, B. 2006. The Human Foot: A Companion to Clinic Studies. Springer-Verlag London Limited.
Kuhn, S.L., Raichlen, D.A., Clark, A.E., 2016. What moves us? How mobility and movement are at the center of human evolution. Evolutionary Anthropology: Issues, News, and Reviews. 25, 86-97.
Latimer, B. and Lovejoy, C.O., 1989. The calcaneus of Australopithecus afarensis and its Implications for the evolution of bipedality. American Journal of Physical Anthropology. 78, 369-386.
Latimer, B. and Lovejoy, C.O., 1990. Metatarsophalangeal joints of Australopithecus afarensis. American Journal of Physical Anthropology. 83, 13-23.
Leakey, M.D. and Hay R.L. 1979. Pliocene footprints in the Laetolil Beds at Laetoli, northern Tanzania. Nature 278, 317323.
Leakey, M.D., 1981. Discoveries at Laetoli in northern Tanzania. Proceedings of the Geologists Association. 92, 81-86.
76


Leakey, M.D. 1987. The Hominid Footprints: Introduction. In: Laetoli: A Pliocene Site in Northern Tanzania. Leakey, M. D. and Harris, J. M., (Editors). Clarendon Press Oxford.
Leakey, M. D. and Harris, J. M., (Editors). 1987. Laetoli: A Pliocene Site in Northern Tanzania. Clarendon Press Oxford.
Leakey, M. G., Spoor, F., Brown, F. H., Gathogo, P. N., Klarle, C., Leakey, L. N., McDougall, I. 2001. New hominin genus from eastern Africa shows diverse middle Pliocene lineages. Nature 410: 433-440.
Lockley, M. G. 1998. The Vertebrate Track Record. Nature 396: 429-433.
Lovejoy, C. O. 1981. The Origin of Man. Science 211 (4480): 341-350.
Lovejoy, C.O., 2009. Reexamining Human Origins in Light of Ardipithecus ramidus. Science. 326, 74+
Lovejoy, C.O., Latimer, B. M., Spurlock, L., Haile-Selassie, Y. 2016. The Pelvic Girdle and Limb Bones of KSD-VP-1/1. In: Haile-Selassie Y., Su D.F., editors. The Postcranial Anatomy of Australopithecus afarensis. Springer Netherlands. 155-178.
Masao, F.T., Ichumbaki, E.B., Cherin, M., Barili, A., Boschian, G., lurino, D.A., Menconero, S., Moggi-Cecchi, J., Manzi, G. 2016. New footprints from Laetoli (Tanzania) provide evidence for marked body size variation in early hominins. eLife. 5, e19568.
Matthews, N. A. 2008. Aerial and Close-Range Photogrammetric Technology: Providing
Resource Documentation, Interpretation, and Preservation. Technical Note 428. U.S. Department of the Interior, Bureau of Land Management, National Operations Center, Denver, Colorado.
Matthews, N., Noble, T., Breithaupt, B. 2016. Close-Range Photogrammetry for 3-D
Ichonology: The Basics of Photogrammetric Ichnology. In: Falkingham, P. L., Marty, D., Richter, A., eds. Dinosaur Tracks: The Next Steps. Indiana University Press.
McHenry, H. M. 1986. The First Bipeds: A Comparison of the A. afarensis and A. africanus Postcranium and Implications for the Evolution of Bipedalism. Journal of Human Evolution 15: 177-191.
McHenry, H.M., 1991. Sexual dimorphism in Australopithecus afarensis. Journal of Human Evolution. 20, 21-32.
McHenry, H. M. 1994. Early Hominid Postcrania Phylogeny and Function. In: Integrative Paths to the Past: Paleoanthropological Advances in Honor of F. Clark Howell. Corruccini, R. S. and Clochon, R. L., editors. Prentice Hall. 250-268.
77


McHenry, H. M., Jones, A.L., 2006. Hallucial convergence in early hominids. Journal of Human Evolution. 50, 534-539.
McHenry, H.M. and Brown, C.C. 2008. Side steps: the erratic pattern of hominin postcranial change through time. Journal of Human Evolution. 55: 639-651.
Meldrum, D. J., Lockley, M. G., Lucas, S. G., Musiba, C. 2011. Ichnotaxonomy of the Laetoli trackways: The earliest hominin footprints. Journal of African Earth Sciences60 1-12.
Musiba, C.M., Tuttle, R.H., Hallgrimsson, B., Webb, D.M., 1997. Swift and sure-footed on the Savanna: A study of Hadzabe gaits and feet in Northern Tanzania. American Journal of Human Biology. 9, 303-321.
Musiba, C. M., Magori, C., Stoller, M., Stein, T., Branting, S., Vogt, M., Tuttle, R., et al. 2007. Taphonomy and paleoecological context of the Upper Laetolil Beds (Localities 8 and 9), Laetoli in northern Tanzania. In: Bobe, R., Alemseged, Z., Behrensmeyer, A. K., eds. Hominin Environments in the East African Pliocene: An Assessment of the Faunal Evidence. Springer Netherlands. 257-278.
Musiba, C. M., Mabula, A. Selvaggio, M. Magori, C. C. 2008. Pliocene Animal Trackways at Laetoli: Research and Conservation Potential. Ichnos 15: 166-178.
Nathan, R., Getz, W. M., Revilla, E., Hoyoak, M., Kadmon, R., Saltz, D., Smouse, P.E. 2008. A movement ecology paradigm for unifying organismal movement research. PNAS 105 (49): 19052-19059.
Pickford, M., 2006. Paleoenvironments, Paleoecology, Adaptations, and the Origins of Bipedalism in Hominidae. 175-198.
Plavcan, J.M., Lockwood, C.A., Kimbel, W.H., Lague, M.R., Harmon, E.H., 2005. Sexual dimorphism in Australopithecus afarensis revisited: How strong is the case for a human-like pattern of dimorphism? Journal of Human Evolution. 48, 313-320.
Potts, R., 1994. Variables versus models of early Pleistocene hominid land use. Journal of Human Evolution. 27, 7-24.
Potts, R., 1998. Variability Selection in Hominid Evolution. Evolutionary Anthropology. 7, 81-96.
Potts, R., 2007. Environmental Hypotheses of Pliocene Hominin Evolution. In: Bobe,
Alemseged, and Behrensmeyer (eds.) Hominin Environments in the East African Pliocene: An Assessment of the Faunal Evidence, 25-49.
Potts, R., 2012. Environmental and Behavioral Evidence Pertaining to the Evolution of Early Homo. Current Anthropology. 53:56, S299-S317.
78


Prang, T.C., 2015. Calcaneal robusticity in Plio-Pleistocene hominins: Implications for locomotor diversity and phylogeny. Journal of Human Evolution. 80, 135-146.
Raichlen, D. A., Pontzer, H., Sockol, M. D. 2008. The Laetoli footprints and early hominin locomotor kinematics. Journal of Human Evolution 54: 112-117.
Raichlen, D. A., Gordon, A. D., Harcourt- Smith, W. E.H., Foster, A. D., Haas Jr., W. R. 2010. Laetoli Footprints Preserve Earliest Direct Evidence of Human-Like Bipedal Biomechanics. PLoS One 5 (3): e9769.
Reed, K.E., 1997. Early hominid evolution and ecological change through the African Plio-Pleistocene. Journal of Human Evolution. 32, 289-322.
Reno, P. L. McCollum, M. A., Meindl, R. S., Lovejoy, C. O. 2010. An enlarged postcranial sample confirms Australopithecus afarensis dimorphism was similar to modern humans. Philosophical Transactions of the Royal Society B 365: 3355-3363.
Richmond, B.G., Jungers, W.L., 1995. Size variation and sexual dimorphism in
Australopithecus afarensis and living hominoids. Journal of Human Evolution. 29, 229-245.
Robbins, L.M. 1987. Hominid footprints from Site G. In: Laetoli: A Pliocene Site in Northern Tanzania. Leakey, M. D. and Harris, J. M., (Editors). Clarendon Press Oxford.
Ruff, C., 2002. Variation in Human Body Size and Shape. Annual Review of Anthropology. 31, 211-232.
Sayers, K. and Lovejoy, C. O., 2008. The Chimpanzee Has No Clothes: A Critical Examination of Pan troglodytes in Models of Human Evolution. Current Anthropology. 49:1, 87-114.
Stern, J. T., and Susman, R. L., 1983. The Locomotor Anatomy of Australopithecus afarensis. American Journal of Physical Anthropology. 60, 279-317.
Su D. F. and Harrison T. 2008. Ecological implications of the relative rarity of fossil hominins at Laetoli. Journal of Human Evolution 55:672-681.
Tattersall, I., 1986. Species recognition in human paleontology. Journal of Human Evolution. 15, 165-175.
Tuttle, R. H., 1987. Kinesiological inferences and evolutionary implications from Laetoli
bipedal trails G-1, G/23, and A. In: Leakey, M. D. and Harris, J. M., Editors. Laetoli: A Pliocene Site in Northern Tanzania. Clarendon Press Oxford. 503-523.
Tuttle, R. H., Webb, D., Weidl, E., Baksh, M. 1990. Further Progress on the Laetoli Trails. Journal of Archaeological Science 17: 347-362.
79


Tuttle, R. H., Webb, D.M., Baksh, M. 1991. Laetoli Toes and Australopithecus afarensis. Human Evolution 6 (3): 193-200.
Tuttle, R. H., Webb, D. M., Tuttle, N. I. 1991. Laetoli Footprint Trails and the Evolution of
Hominid Bipedalism. Origine(s) De La Bipedie Chez Les Hominides. Paris: Cahiers de Paleioanthropologie, Editions du CNRS : 203-221.
Tuttle, R. H. 2008. Footprint Clues in Hominid Evolution and Forensics: Lessons and Limitations. Ichnos. 15, 158-165.
Ward, C. V., 2002. Interpreting the Posture and Locomotion of Australopithecus afarensis: Where Do We Stand? Yearbook of Physical Anthropology. 45, 185-215.
Ward, C. V., Kimbel, W. H., Harmon, E. H., Johanson, D.C., 2012. New postcranial fossils of Australopithecus afarensis from Hadar, Ethiopia (1990-2007). Journal of Human Evolution. 63, 1-51.
Ward, C.V., Kimbel, W.H., Johanson, D.C., 2011. Complete Fourth Metatarsal and Arches in the Foot of Australopithecus afarensis. Science. 331, 750-753.
Ward, C.V., 2013. Postural and Locomotor Adaptations of Australopithecus Species. In: Reed, K.E., Fleagle, J.G., Leakey, R.E. (Eds.), The Paleobiology of Australopithecus, Vertebrate Paleobiology and Paleoanthropology. Springer Netherlands. 235-245.
White, T.D. and Suwa, G., 1987. Hominid Footprints at Laetoli: Facts and Interpretations. American Journal of Physical Anthropology. 72, 485-514.
Zaitsez, A. N., Wenzel, T., Spratt, J., Williams, T. C., Strekopytov, S., Sharygin, V. V., et al. 2011. Was Sadiman volcano a source for the Laetoli Footprint Tuff? Journal of Human Evolution S'! : 121-124.
Zipfel, B., DeSilva, J. M., Kidd, R. S., Carlson, K. J., Churchill, S. E., Berger, L. R. 2011. The Foot and Ankle of Australopithecus sediba. Science 333: 1417-1420.
80


Full Text

PAGE 1

i A C OMPARATIVE A NALYSIS OF N EWLY D ISCOVERED PLIOCENE H OMININ F OOTPRINTS FROM LAETOLI, TANZANIA by Alex J. Pelissero B.A., University of Colorado Boulder, 2010 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Arts Anthropology Program 2017

PAGE 2

ii 2017 ALEX J. PELISSERO ALL RIGHTS RESERVED

PAGE 3

iii This thesis for the Master of A rts degree by Alex Joel Pelissero has been approved for the Anthropology Program by Charles Musiba, Chair Anna Warrener Neffra Matthews April 28, 2017

PAGE 4

iv Pelissero, Alex J. (M. A. Anthropology) A Comparative Analysis of Newly Discovered Pliocene Hominin Foo tprints from Laetoli Thesi s directed by Associate P rofessor Charles M. Musiba ABSTRACT The discovery of the 3.6 million year old Laetoli footprints was one of the most significant advances in our understanding of human evolution. In ad dition to confirming deep roots of bipedality as the defining characteristic of the hominin clade, its association with the contemporaneous species Australopithecus afarensis demonstrated that bipedality emerged well before encephalization in human evolution. Preserved trackways, like those seen at Laetoli, provide a crucial glimpse at both early hominin mobility and its paleoecological context. The recent discovery of additi onal, markedly larger, contemporaneous hominin footprints southwest of the Site G hominin footprints provides a critical comparative data set to test hypotheses about Pliocene hominin locomotion and ichnotaxonomy. Utilizing photogrammetric imagery and mode ling, this thesis aims to compare the morphology of the two sets of trackways, as well as the gait patterns of the individuals responsible for creating the trackways, alongside modern human data. There are several affinities both morphologically and taphon omically between these two trackways, but the size differences seen between these two sets of trackways is likely associated with great differences in body size. This could be indicative of either significant size dimorphism and different locomotor reperto ires in Au. afarensis, or potentially the presence of another hominin species at Laetoli. Situated within the broader paleoecological context of the Laetoli site, the qualitative and quantitative results of this analysis provides further insights into ques tions of early hominin taxonomic diversity and the evolution of upright bipedal locomotion. The form and content of this abstract are approved. I recommend its publication. Charles Musiba

PAGE 5

v TABLE OF CONTENTS CHAPTER I. INTRODUCTION ............... ......................................................................................1 II. BACKGROUND RESEARCH AND CURRENT LITERATURE ...............................3 Theoretical Background ............................................................... ....................3 Current Literature .............................................................. ...............................6 Geology and Paleoecology of Laetoli .................... ...............................6 Locomotion of Laetoli Tra ck makers ............................. ..................... 12 Locomotion of Australopithecus afarensis ..................... ..................... 16 Pliocene Hominin Adaptive Radi ation and Australopithecine Locomotor Variation .................... ......................... ..............................22 III. MATERIALS AND METHODS ............................................... ........................... ...29 Materials and Data Collection .......................................... ..............................29 Sample ................................................................. ..............................29 Descriptions of the Site S Hominin Footprints ...... ..............................30 Photogrammetric Methods... ....... ......................... ..............................37 Comparative Analysis ......................................... ............................... 46 Metric Comparison and Analysis............................................ 4 7 Procrust es Landmark Shape Analysis ...... ..............................48 IV. RESULTS ............................................................................... ................ ..............53 Comparison with other hominins and modern humans .... ............. .................53 Kernel Density Plot and Procrustes Landmark PCA ........ ..............................59 Summary of Results ......................................................... ..............................62

PAGE 6

vi V. DISCUSSION ...................... .................................................... ..............................63 VI. CONCLUSION ....................................................................... .. ............................71 REFERENCES ........................................... ............................................................... ............ 73

PAGE 7

vii LIST OF TABLES TABLE 3.1 Photogrammetric Capture Data.......................................................................................39 3.2 Landmark P lacement .......... ............................................................... ..............................49 4.1 Mean Metric Results...................................................................................................... ..53

PAGE 8

viii LIST OF FIGURES FIGURE 2.1 Stratigraphy of Laetolil Beds ............................................................... ...............................7 2.2 Map of the Northern Tanzania Highlands.......................................... ...............................9 2.3 Maps of Laetoli Region ................................................................. ................................... 10 2.4 Map of the Laetoli Site........................................................................ .............................10 2.5 A L 3 33 15 p artial foo t . . . . . . . . . . . . 17 2.6 Burtele partial foot .............................................................................. ..............................2 1 3.1 TP2 Trench Footprint s ......................................................................... ............................33 3.2 M9 Trench Footprints......................................................................... ............................ ..34 3.3 L8 Trench Footprints.......................................................................... ..............................35 3.4 L8 Trench Footprints.......................................................................... ...................... ........36 3.5 G Track Photogrammetric Model....................................................... ..............................41 3.6 M9 Photogrammetric Model............................................................... ..............................42 3.7 TP2 Ph otogrammetric Model........................................................................................... 43 3.8 L8 Photogrammetric Model................................................................ ..............................44 3.9 Meas urement Schemes..................................................................... ..............................47 3.10 S 1 Prints Used in 2 D Landmark Analysis...................................... ..............................51 3.11 Site G Sample used in 2 D Landmark Analysis................................................... ....... ...52 4.1 Footprint Length Comparative Box plots.........................................................................54 4.2 Step length Comparative Box plots............ .....................................................................55 4.3 Stature vs. Foot Length Plot .............................................................. ..............................56 4.4 Metric Data PCA Plot ................................. .............................................................. ....... .57 4.5 DFA Plot ..........................................................................................................................58 4.6 Landma rk PCA Plot......................... ................................................................................59

PAGE 9

ix 4.7 S 1 Print Kernel Density Plot .................................................................................... .......60 4.8 G Avg. Kernel Density Plot ................ ................................................ .............................. 60 4.9 2 D PCA Relative Deformations ........................................................ ..............................61 4. 10 2 D PCA Overall Relative Deformation ............................................ ............................ .62

PAGE 10

1 CHAPTER I INTRODUCTION In the northern highlands of Tanzania, nestled within an uplift of the East Afri ca n Rift system, and surrounded by innumerable eroding fossili ferous deposits and outcrops, is one of the most important and spectacular paleoanthropological finds in the world: the Laetoli Site G hominin footprint trail s Imprinted in volcanic ash 3.66 mi llion years ago by Pliocene hominins, the Site G trackway, discovered in 1978 by a team led by Mary Leakey upended the scientific community and was a crucial advance in our understanding of human evolution (Leakey and Hays 1987). In addition to their rem ark able state of preservation, the footprints confirmed the deep roots of bipedality in hominin evolution, and their association with the then recently described Australopithecus afarensis demonstrated the emergence of the suite of bipedal locomotor adap tations before encephalization in the evolution of our clade. The trackways have remained a continuing source of important research and controversy, even four decades later. The three track makers' locomotor capabilities, gait, body size, pedal morphology, social grouping, and paleohabitat have all been intensely scrutinized and debated, and the continued lack of hominin pedal fossils found at Laetoli casts a shadow over many analyses. The lack of an y almost complete associated foot elements at H adar, where the vast majority of Au. afarensis specimens have been discovered, is similarly frustr ating. While scarcely a season has passed without field research at the 20km Laetoli site at large, and despite numerous erosional and excavated exposures o f associated animal trackways, it wasn't until 2015 that additional hominin prints were discovered. Preserved in the same geological horizon and located only 150 meters to the southwest of the G trackway (Masao, et al. 2016), the markedly larger Site S foo tprints have the potential to reinvigorate hominin locomotor research and stoke new questions about the hominins the inhabited the Pliocene Laetoli landscape.

PAGE 11

2 While debates regarding speciation within the hominin clade and intra specific variation have long been a part of the paleoanthropological dialogue, in the past two decades researchers have to contend with both an increasingly complex and diverse hominin evolutionary tree particularly the marked morphological variation in areas such as locomotor r epertoire, body proport ions, and masticatory complexes. The discovery of the Site S footprints pro vides an important illustration of the emerging picture of hominin variation, both within and between species. With this in mind, in order to situate these new ly discovered footprints within (or agai nst) our current understandings of hominin evolution and behavior, comparative analyses must be undertaken with other, similar datasets including the Site G trackway and habitually unshod modern humans While there are numerous taphonomic, substrate, and biomechanical variables to consider in footp rint formation, they provide us with the best direct evidence of how hominins moved, and any notable differences that can be discerned between different populations is pot entially illuminating of morphological and/or locomotor variation. Some specific aims of this thesis include: 1. Direct comparison, qualitatively and quantitatively, of the three currently known sections of the Site S trackway with the Site G trackway via us e of 3 D photogrammetry. 2. Alongside the direct visual qualitative and quantitative comparison of these two sets of trackways, additional comparison with unshod mode rn human data utilizing footprint measurement and gait metric data and a discriminant functi on analysis 3. Determine if the footprint morphology of the Site S prints demonstrates significant differences from those of Site G individuals via Procrustes fitted landmark principal components analysis.

PAGE 12

3 CHAPTER II BACKGROUND RESEARCH AND CURRENT LITER ATURE Theoretical Background In studies such as this, it is important to contextualize the research within our current understanding of the evolution of biped ality in the hominin clade and within model s of hominin behavior and mobility. One of the longest standing hypotheses regarding the evolution of upright posture and bipedal locomotion posits that it emerged as an adaptation to the opening of the East African landscape from primarily forested to savannah grasslands. However, there is a growing amount of evidence from early hominin sites that indicate many of the earliest known definitive bipeds were inhabiting environments that wer e still largely wooded or mosaic (Potts 1998; Behrensmeyer and Reed 2013). The existence of early putativ e Miocene hominin Ardipithecus in what is argued to be a closed woodland environment, and Sahelanthropus in what is interpreted as a more open habitat demonstrates t he variable nature of early homin in paleohabitats (Lovejoy, 2009 ; Brunet, et al. 2002 ). Man y of the models, such as Wheeler's (1991) thermal regulation hypothesis, that centered around open grasslands as the driving force of bipedality, appear to be teleological and concerned with function over adaptation and behavioral response. Nevertheless, t here is a notable expansion of African grasslands within the Pliocene, and as Dominiguez Rodrigo (2014 ) points out, savannah as a biome encompasses a variety of habitats, including closed woodland, which entail a variety of adaptive responses. The mosaic n ature of many Miocene and early Pliocene African environments largely precludes open grasslands as the driving force of bipedality, as does climatic evidence that shows aridification and a marked increase in continent wide grassland expansion not occurring until late r in the Pliocene, around 3.0 2.8 mill ion years ago (Behrensmeyer 2006 ). Thus, other selective pressures had to be acting upon the earliest hominins to push bipedality as

PAGE 13

4 an adaptive response after its initial emergence in the Miocene. Behaviora l and adaptive flexibility in the face of climatic and environmental uncertainty is often posited, which also ties into diet and subsistence behavior, and cannot be extracted from discussions regarding m obility (Dominiguez Rodrigo 2014 ; Kuhn, et al. 2016). While external ecological factors are crucial to understanding hominin adaptation, their behavioral responses, both on an individual and on a population level are also necessary to understand. Behavioral ecology models have a complicated relationship w ith anthropology, yet they remain the best framework with which to understand early hominin behavior and mobility, particularly prior to the earliest archaeological evidence. In order to understand what drove bipedality as an adaptive response, regardless of the specific ecological context, one needs to approach what drives hominin mobility from a theoretical perspective. Nathan, et al. (2008), in their movement ecology framework, refer to this as the "internal state" of movement. While constrained by the c apabilities of the organism itself and the external factors, the internal state of movement cuts across theoretical models based on behavioral ecology. This relationship between internal and external dynamics shapes mobility throughout human evolution, fro m the australopithecines to Paleolithic hunter gatherers and onward. One of the most basic driving urges is subsistence, and as Kuhn, et al. (2016) state, "variation in diet directly implicates movement." For early hominins, habitat and dietary variation c an theoretically be correlated to differences in subsistence behavior, and potentially, locomotor adaptation, which in turn, could be linked to adaptive radiation. The variable presence and retention of arboreal traits in australopithecines well into the P leistocene and the emergence of Homo suggest behavioral variability reflecting adaptations to different habitats (Ward 2002). While some models of the evolution of bipedality have explicitly linked its emergence to subsistence and provisioning behaviors (L ovejoy 1981; Hunt 1994), it is likely that dietary variables were coupled with other behavioral changes and adaptations alongside the e volution of bipedal locomotion.

PAGE 14

5 With both extant primate and modern human behavioral models, the issues with inferring early hominin behavioral ecology stem from the general problems with the use of analogy within the paleontological and archaeological record. Behavior does not necess arily preserve directly in either r ecord so researchers must utilize the evidence at thei r disposal, namely skeletal morphologies, artifacts, and associated faunal remains, to infer it indirectly. For early hominins, the issue lays in the fine line between over anthropomorphizing homin in behavior and drawing too much from the behavior of our closest evolutionary relatives, who continue to occupy the traditional niche of apes, and have undergone their own evolutionary trajectory. In both instances, despite the issues, there is merit i n some of the perspectives that behavioral ecology can offer studies of early hominin mobility. Humans and early hominins alike, are nevertheless primates, and extant species, and Pan in particular, offer the easiest model of social behavior for comparison (Foley and Gamble 2009). While other animals may practice co ordinated hunting, food sharing, tool use, etc (Sayers an d Lovejoy 2008), other primates and all the different ecologic al and social adaptations they possess are the best comparative model currently at our disposal. This is true with not only adaptive be havior, but also with morph ological adaption and radiation. This is best exemplified by the frequent existence of more than one primate species in a given habitat, which is seen in both the fossil record and in modern day primate communities.

PAGE 15

6 Current Literature Geology and Paleoecology of Laetoli The geology of the Laetoli area has been extensively mapped and studied, especially since extensive research began at the site in the 1970s. The Plio Pleisto cene sedimentary and volcanic deposits (Fig 2.1) which lay directly upon Precambrian metamorphic basement rocks, consist of a series of beds spanning from 4.36 2.3ma: Lower Laetolil (~4.36ma 3.85ma), Upper Laetolil (~3.85ma 3.63 ma), Ndolanya (~3.58 2.66ma ), and the Naibadad Beds (~2.15ma 2.06ma). The Upper Laetolil beds and the younger Ndolanya beds are separated by a distinctive yellow marker tuff throughout much of the site, likewise with the Ogal lavas separating the Ndolanya Beds from the upper deposits. The Upper Laetolil bed, the l argest and most fossiliferous of the deposits, and the sequence that contains the footprint tuff, is largely comprised of aeolian and air fall/fall out tuffs, which, due to their geographic extent at the Laeto li site, have undergone extensive, but variable fluvial and standing water reworking at the different localities (Hay 1987; Musiba, et al. 2007; Ditchfield and Harrison 2011). The major fossiliferous tuffs (Tuff 6 through Tuff 8, largely) are dominated by volcanic ash fall out deposits. Tuff 7 of which the lower portion has been dubbed the footprint tuff, is the thickest and most distinctive (Ditchfield and Harrison 2011). The footprint tuff consists of fine grained, heavily compacted carbonatite ash, which varies in thickness from 12 15cm, and is subdivided into two differing lithological layers, which in turn document numerous separate eruption and ash fall events, many of which contain the eponymous footprints (Hay 1987). The frequency of the layers, and the excellent preservation of the footprin ts and very existence and high preservation of ra indrop imprints points towards very short depositional period s likely only a few weeks (Hay 1987).

PAGE 16

7 Figure 2.1: Stratigraphic sequence at Laetoli (left) (Harrison 2011) and microstratigraphy of the fo otprint tuff (right) (Hay 1987). This has significant implications for studying the paleoecological makeup of the Laetoli site during this period, an opportunity that most time averaged fossil deposits cannot provide. The greater Laetoli/Eyasi Plateau r egion (Fig. 2.2) is dotted with ancient volcanoes that are responsible for the ash fall tuffs, although it is generally believed that the volcano Satiman is the likely source of the ash fall deposits making up the footprint tuff, based on its age and the g eochemical composition of the rocks, sediments, and lava flows found at the extinct volcano (Hay 1987; Meldrum et al. 2011). Although it is not a definitive relationship, as depositional and temporal factors have resulted in numerous geologic discrepancies (Meldrum, et al. 2011; Zaitsev, et al. 2011), Satiman currently remains the most plausible

PAGE 17

8 candidate for creation of the Lae tolil beds. The preservation of the animal trackways in the ash fall tuffs is a crucial piece of evidence for the paleoecological construction of Pliocene Laetoli, given the relative dearth of fossilized remains at the site (Su and Harrison 2008) The discovery of the Site G hominin trackway and its central role in understanding the evolution of bipedality has given constructing the paleoecology of Pliocene Laetoli habitats accurately an added importance. In the initial monograph published on the s ite, Leakey (1987) argued for the paleoecological interpretation of a largely dry open grassland biome. This interpretation was supported by their analysis of the faunal evidence accumulated at the time, and possibly influenced by the then prevailing "open savannah" hypothesis for the emergence of hominin bipedality. Since the initial publications of research emerging from the Laetoli site, there has consistently been competing interpretations of the paleoecological evidence with taphonomic and sampling biases argued to be the culprit for earlier "Serengeti like" interpretations (Musiba, et al. 2007) There is mounting evidence for a greate r occurrence of both closed and open wooded ha bitats, a trend that, as noted above, is becoming broadly seen across many late Miocene and Pliocene hominin sites (Behrensmeyer and Reed 2013). Faunal, paleofloral, dietary isotopic, and pollen evidence accumu lated over the last three decades of research at Laetoli support a largely wooded landscape with a mix of more open grassland habit ats (Kingston and Harrison 2007; Harrison 2011; Su 2011). Other recent interpretations agree with this general assessment, but argue that the paleoenvironment of Pliocene Laetoli was much more variable and mosaic in its habitat makeup, likely containing starker divisions in its biomes than seen presently, due to highly variable local climatic factors (Musiba, et al. 2007; Andrews and Bamford 2008 ; Musiba et al. 2008). Analyses focusing on the past and current vegetation landscape, including fossil remains and isotopic signals at Laetoli note many similar biomes, but a distinct difference in the overall make up at the site over time. Both Andrews and Bamford (2008) and

PAGE 18

9 Figure 2.2: The Northern Highlands of Tanzania, showing paleo volcanoes and notable sites. Volcano (1) is Satiman. From Barboni (2014).

PAGE 19

10 Figure 2.3: Map of the Laeto li region at large, from Harrison ( 2011 ) Figure 2.4: Map of the Laetoli Site. Localities are designated by numbers, footprint sites by letters. Modified f rom Musiba, et al. (2008)

PAGE 20

11 Barboni (2014) note the existence of C4 grasslands in Pliocene Laetoli, yet argue that they only make up a smaller portion of the vegetative biomass at the site compared to C3 plant material. However, Barboni (2014) notes that in the phytolith evidence, a steady expansion in C4 grasses relative to woody cover can be seen after the deposition of Tuff 6 (~3.7Ma). Skewing towards the more wooded end of the biome spectrum, Pliocene Laetoli would likely have been a patchwork of woodland, shrub lands, and grasslands of varying compositions and more heavily wooded than the present day L aetoli region (Harrison 2011). This has important implications for subsequent interpretations of the locomotor adaptations of the Laetoli hominins, as well as other crucibles of bipedal evolution in Africa. The effects of the volcanic and te ctonic activity on the landscape and paleoecological makeup at Laetoli during the mid Pliocene also cannot be discounted. Both Andrews and Bamford (2008 ) and Barboni (2014) both note how the processes of uplift resulting in the creation of the East African rift would have created variable rainfall in Pliocene Laetoli, which arguably contributed to a stark mosaic like habitat structure. The frequent disruptions from volcanic activity and ash fall undoubtedly affected both the floral and faunal communities, although to what extent is un clear. Andrews and Bamford (2008 ) argue that much as present volcanic and tectonic activity affects the vegetation structure on either side of the East African R ift, it would be have particularly notable in the Pliocene, contr ibuting to habitat variability. Ditchfield and Harrison (2011) argue that the effects of the periodic ash falls would be incredibly disruptive for woodlands, resulting in periods with higher amounts of open grasslands. However, they argue that these disru ptions would be relatively short lived, and the paleoenvironmental balance of Pliocene Laetoli would rebound as soils redeveloped to allow for woody growth. A frequently shifting landscape and vegetation structure seems to be the ecological settings in w hich the Laetoli hominins inhabited, and it likely had an effect on their adaptive responses

PAGE 21

12 Locomot ion of the Laetoli Track makers There are several notable features that separate the morphology of the modern human foot from that of the rest of the ape s, all of which are functionally important in our obligate bipedal locomotion. Alongside our comparatively shortened toes, which decrease in size from the first digit, one of the most noticeable features is our robust, non grasping hallux. The reorganizati on of the forefoot allows it to function as an effective lever during bipedal gait, with the robust hallux playing a key role in propelling the body forward. In addition to their abductable hallux, non human primates also lack the derived medial longitudin al arch, possessing only the transverse arch of the foot. This gives their feet (and as a result, their footprints) a flattened, splayed appearance, with a noticeably large gap between their first and second digits. The medial longitudinal arch is a signif icant evolutionary development for obligate bipedalism, allowing for greater stability, shock absorption, and medial weight transfer (Aiello and Dean 1990). While arch formation and structure is variable in modern humans, it remains a key difference separa ting human like foot function from less derived forms, even among modern humans with non pathological flat feet (Klenerman and Wood 2008). Other key shifts in modern foot construction include re organization of the ankle joint and a relatively stiff mid fo ot, as arboreal adapted apes have a great deal of mid foot flexibility, known as the mid tarsal break (Aiello and Dean 1990). While there are limits to the inferences that can made on footprints, they nevertheless can provide insights into certain skeleta l and soft tissue morphologies and gait patterns by looking at their construction and pressure distribution (Bennet and Morse 2014). However, as the Laetoli prints have demonstrated, interpreting these subtle signals can be contentious. The Site G footpr ints are undeniably made by a group of habitually upright bipedal individuals, a fact recognized immediately upon their discovery (Leakey and Hay 1979) However, debates within the paleoanthropological community have largely centered around

PAGE 22

13 the relative hu man like nature of the bipedality seen in the footprints, as wel l as what hominin species were responsible for their creation. Upon their discovery in 1978, only two individuals were initially recognized (Leakey a n d Hay 1979), although as additional, bette r preserved prints were discovered, it became clear that the second, larger, tracks were in fact the prints of a third individual overlaying those of the second (Leakey 1987). More recent analysis of the overlaying prints has provided glimpse of a potentia l fourth individual in G 2/3 based on the presence of what appears to be additional halluc al imprints (Bennet, et al. 2016a). However, this could also be a false signal from the trackway cast, as White and Suwa (1987) remark that lagomorph prints are abundant in the footprint tuff, and note the presence of a track crossing the G 1 36 print. The overlaying of these two (or three) individuals has complicated analysis and comparison of the two parallel trackways, as the prints of the G 3 individual largel y obliterated the posterior portion of the G 2 prints, preventing accurate measurement of the true length of the G 2 individual's footprints (Tuttle 1987). While some prints of t he G 2/3 trackway allow for better estimation of the shape and metrics of the G 2 individual, most researchers have opted to focus their analyses on G 1 and G 3, the two smaller of the three individuals. While preservation of the prints varies wildly along the trackways (particularly in the northern end), the G 1 prints themselves are noticeably better preserved than the G 2/3 trail, so much that Leakey (1979) described them as likely having been imprinted on a more dense, compacted surface compared to the other trail, suggesting that the individuals walked through the area at diff erent times within the same ash fall event. However, White and Suwa (1987) remarked that the different walking surface hypothesis had been largely abandoned by the original researchers after further study of the site. Additionally, in their analyses of th e Site G Trackway, both Robbins (1987) and Tuttle (1987) remark that it is likely that all three individuals were walking through concurrently, and the G 1 and G 3 individual were attempting to keep pace with the G 2 individual, based on how closely the

PAGE 23

14 pr ints mirror each other in placement. Tuttle (1990), nevertheless states that temporally separ ate footprint creation events are still likely and cannot be definitively fully ruled out. In addition to ar guing for the presence of a medial longitudinal arch both Robbins (1987) and Tuttle (1987) remark on the affinities the tracks share with modern human locomotor function, particularly in the heel, ball, and toes. Weight transfer from heel to toe along the lateral side of the arch, followed by a medial sh if t towards the adducted hallux and its notably deep impression, suggesting toe off prior to swing phase, is noted by both authors. Tuttle (1987) goes as far to say that the prints are "indistinguishable from those of modern humans, not noting any phalange al curvature or halluc al gaps outside of modern human variation, although their stride lengths were noticeably shorter than seen in modern humans. White and Suwa (1987) dispute this observation by Tuttle (1987), but nevertheless state that they are signifi cantly more similar to modern human footprints than any other known primate. Most importantly, they state that the foot of the track makers need not be identical to that of H. sapiens to perform human like bipedal locomotion. In their comparative analysis utilizing footprints of the habitually unshod Machiguenga peoples, Tuttle, et al. (1990) reaffirm their argument as to the remarkable human like nature of the prints. Stern and Susman (1983) disputed the assertions of Leakey (1987), Robbins (1987), Tuttle (1987), and White and Suwa (1987), arguing that the Laetoli footprints resemble a much more transitional form, sharing affinities with chimpanzee footprints and displaying poor ly developed morphologies that are present in modern humans. Tuttle, et al. (199 0) aggressively disputed the assertions made by Stern and Susman (1983), claiming that they were built upon incomplete and incorrect data and analysis. Further comparative studies with the Hadzabe of Tanzania by Musiba, et al. (1997) bolstered the argument that the Laetoli prints were remarkably human like in form and function. Given the differences in stature, walking speed and stride length between the Hadza and Machiguenga peoples (Musiba, et al. 1997), the affinities seen in the Laetoli prints are notab le.

PAGE 24

15 Nevertheless, the debates regarding the nature of the Laetoli track makers has continued into the 21st century. A more recent statistical study by Berge, et al. (2006) reported some affinities with chimpanzee gait and morphology such as foot roll of f at stance phase and pedal proportions. However, their analysis overwhelmingly supports the notion that the Laetoli track makers displayed more human like traits in their pedal function than those of a chimpanzee, particularly in the anterior part of the foot (Berge, et a l. 2006). Raichlen, et al. (2008 ) studied the locomotor biomechanics of the Laetoli prints, and determined that the track makers could have either used modern human like extended hind limb gait or more chimpanzee like "bent knee bent hip" gait to create the footprints. However, in a later study utilizing three dimensional laser scan modeling of pressure kinematics, Raichlen, et al. (2010) determined that the Laetoli track makers walked with extended limb gait, as the prints displayed signif icantly more morphological similarities with experimental modern human prints walking normally than those made by modern humans walking with bent knee bent hip gait. A pedobarographic statistical mapping analysis conducted by Crompton, et al. (2012) reache d similar conclusions regarding the human like locomotion of the Laetoli track makers. Bennet, et al. (2016b) and Hatala, et al. (2016) most recently have offered competing conclusions regarding the function of the Laetoli track makers feet. Whereas Ben net, et al. (2016b) maintain the belief that the Laetoli tracks are functionally indistinguishable from modern humans, and promote the idea that hominin foot functionality has possibly remained in evolutionar y homeostasis for the last 3.66M a, Hatala, et al (2016) argue that the Laetoli tracks demonstrate a form of locomotor mechanics unlike humans or chimpanzees, largely based on differences in medial foot rotation and knee and hip flexion during gait. Meldrum, et al. (2011) similarly distinguish the Laet oli track makers from modern human morphology, particularly with what they interpret as evidence of a midfoot flexibility. The mid tarsal break, which is the ability to lift the heel independently from the rest of the foot, is often interpreted

PAGE 25

16 as a primit ive condition absent in modern humans, and prominent in other primates, including chimpanzees (DeSilva 2010; Meldrum, et al. 2011). Greater midfoot flexibility results in noticeably different morphological signals in footprint formation, particularly in pr essure dynamics. Meldrum, et al. (2011) interpret several of the Site G tracks as displaying evidence of greater midfoot flexibility, as well as the lack of a medial longitudinal arch. While they do not dispute the clear adapations for terrestrial bipedali ty seen in the Site G footprints, such as the shortened toes and the adducted hallux, Meldrum, et al. (2011) assert that the Laetoli track makers' midfoot flexibility distinguishes them from modern human locomotion. However, Meldrum, et al.'s (2011) interp retation is based on the assumption that the mid tarsal break occurs at the calcaneocuboid joint, which correlates with skeletal reconstruction overlays on the G 1 prints. DeSilva (2010) disputes this long held interpretation of the mid tarsal breaks, and argues for its placement more anteriorly, at the cuboid metatarsal joint. Additionally, DeSilva, et al. (2015) argue that midfoot flexibility is present in varying degrees in modern humans, and is much more complex of a feature than previously realized, an d does not necessarily disprove the existence of a medial longitudinal arch. Bennet and Morse (2014) also dispute the mid tarsal break interpretation, and posit substrate and gait mechanic factors for footprint morphologies argued by Meldrum, et al. (2011) Locomotion of Australopithecus afarensis Deeply intertwined with the debate regarding the functional signals interpreted from the Site G trails is whether Au. afarensis is the species responsible for creating the footprints. The only presently known fossilized hominin remains at Laetoli contemporaneous with the trackways are assigned to Au. afarensis, yet the collection is relatively diminutive compared to the amount of material from Hadar, Ethiopia assigned to the Au. afarensis hypodigm. Furthermore there have not yet been any hominin pedal remains recovered

PAGE 26

17 from Laetoli, while the Hadar site posses nearly the entirety of the Au. afarensis pedal collection, none of which c omprises a complete enough foot to make direct inferenc es on the Laetoli trac ks. Taphonomic factors can be held accountable, as pedal fossils are rare, par ticularly more comp l ete specimens such as the OH 8 foot, for a variety of reasons, not the least of which is their small size and likelihood of preservability. At Laetoli the ta phonomic conditions and paleoenvironment have resulted in a generally smaller fossil collection compared to many fossil localities such as Hadar or nearby Olduvai Gorge, likely due to its lack of permanent lacustrine or riverine water sources (Harrison 201 1). Hominins are a rarity amongst fossil assemblages, but it is particularly so at Laetoli, which may the result of the aforementioned taphonomic bias, or potentially habitat preferences of the sp ecies (Su and Harrison 2008). Figure 2.5 : AL 333 115 partial Au. afarensis foot from Hadar, Ethiopia While numerous pedal element s attributed to Au. afarensis dated from 3.4Ma 3.0M a, have been discovered at Hadar, particularly from the bountiful AL 333 locality, there has only been one associated assemblage of foo t elements, AL 333 115 (Fig. 2.5 ). This

PAGE 27

18 largely fragmentary collection consists of five proximal phalanges, two middle phalanges, and five metatarsal heads, which while informative, lacks the more robust elements that typically preserve an d can be used for comparative analysis (White and Suwa 1987). Other than the small collection from AL 288 1 ("Lucy"), consisting of a talus and two phalanx fragments, the rest of the Hadar pedal collection consists of isolated specimens (Johanson, et al. 1 982; Ward, et al. 2012). Taken in full, numerous similarities can be seen across the combined Hadar collection, as well as a fe w key differences in the tarsal specimens. The metatarsals, while stil l curved, are only slightly more so than modern humans, a nd display distinct dorsal doming of the heads, which separates them from the condition seen in extant apes The dorsoplantarly expanded bases and sulcus separating the metatarsal heads from the epicondyles resembles the morphology of modern human metatars als (Ward, et al. 2012). The phalanges are longer and more curved than in Homo but less so than in modern a pes, and comparable to Ar. ramidus. Similarly, the metatarsal articular facets of the phalanges are more dorsally inclined, at the lower end of what is seen in modern human variation, which is argued to reflect extension during the terminal stance phase of bipedal gait (Stern and Susman 1983; Ward, et al. 2012). The navicular of Au. afarensis has been a particularly controversial element, because o f its large tubero sity and flattened appearance, a primitive condition which has been argued as evidence for its role in weight b e aring during gait, which would possibly presuppose the existence of a medial longitudinal arch (Stern and Susman 1983). Howev er, in addition to the expanded knowledge of metatarsal morphologies that have resulted from continued recovery at Hadar, newer navicular specimens show an expanded area for the insertion of the plantar cubonavicular ligament and a groove for calcaneonavic ular ligaments, as seen in modern humans, which challenges this interpretation (Ward, et al. 2012).

PAGE 28

19 The calcaneal specimens have been similarly controversial, due to their importance in the striking phase of bipedal gait. While the calcanei attributed t o Au. afarensis retain the oval shaped appearance seen in extant apes as well as the "massive" peroneal trochlea and flattened posterior talar surface, they are substantially increased in robusticity in comparison, and are markedly more human like in both cross section and in tuberosity volume (Latimer and Lovejoy 1989). Additionally, Latimer and Lovejoy (1989) disagree with the assessment of Stern and Susman (1983), and argue for the presence of a lateral plantar process in the Au. afarensis calcanei, whic h is a distinct trait seen in later hominin s and modern humans. The tali of Au. afarensis have also been a p oint of contention, as there are marked differences between that of Lucy and the larger specimens such as AL 333 147. Previous analyses pointed t o the more primitive aspects of Lucy's talus, such as the tightly curved, deeply grooved trochlea, and the anterior extension of the fibular malleolar facet on the lateral side of the talar neck, as evidence of arboreal retentions in the species (Susman an d Stern 1983). Howeve r, the more recently discovered larger specimens lack these morphologies, and appear to be in the lower range of modern human variation. These tali exhibit all the morphologies associated with habitual bipedality, including a vertical ly oriented shank an d groove for the flexor halluc u s longus muscle, and a human like axis of rotation at the talocrural joint (Ward, et al. 2012). The differences seen between several of the pedal specimens attributed to Au. afarensis have been attributed by several of the researchers at Hadar to differences in body size and sexual dimorphism (White and Suwa 1987; Ward, et al. 2012) although the potential for different modes of bipedality is often alluded to (Stern and Susman 1983; Harcourt Smith and Aiello 2004). The composite nature of most analyses regarding the foot of Au. afarensis has exacerbated this issue While few researche rs now dispute the terrestrial bipedality of Au. afarensis reconciling the morphology of the Site G footprints with the skeletal anatomy seen in the

PAGE 29

20 Hadar pedal assemblage has prompted debate amongst researchers (Kimbel and Delezene 2009). Tuttle (1987; 2 008) has maintained that t he morphologies exhibited by the Hadar specimens of Au. afarensis would exclude the species from having made the Site G footprints. In addition to the potentially weight bearing naviculars which co uld contradict the e vidence of a medial longitudinal arch in the Site G prints, Tuttle, et al. (1990; 1991a), focus on the disconnect in phalangeal curvature. Whereas the Site G prints display shortened toes statis tically indistinguishable from a Machiguenga footprint sample, and faint lateral toe impressions, Tuttle, et al. (1990; 1991a) argue that the length and level of curvature seen in the Hadar phalanges would have prevented them from making such footprints. Harcourt Smith and Aiello (2004) also argue that the enlarged tuberosity o f the Au. afarensis navicular is evidence of its weight bearing nature, and likel y a lack of a human like arch. Many of the Hadar researchers maintain that the morphologies of the pedal elements, while retaining some primitive morphologies in certain areas are nevertheless capable of having created the prints at Laetoli (White and Suwa 1987; Ward 2002). To support their assertions, White and Suwa (1987) reconstructed an Au. afarensis foot scaled to the size of the G 1 track s from a comp osite of AL 333 elem ents a nd the OH 8 midfoot. Stating that their reconstruction does not definitively prove that Au. afarensis was the track maker, they argue that the species nevertheless cannot be excluded. Ward (2002; 2013), in addition to agreeing that Au. afarensis was the likely track maker at Laetoli, asserts that the body plan and lower limb function of australopithecines emerged within the Au. anamensis/afarensis lineage and was maintained without much variation for at least 700,000 years. Alternatively, Stern and Susman (1983) argued that the transitional morphology of the Au. afarensis foot matched well with the transitional morphology they argued to see in the Laetoli trackways. Tuttle (1990) disputed the arguments made by White and Suwa (1987), specifically citi ng their use of the OH 8 foot, which at roughly 1.75ma, is significantly younger than any of the Hadar Au. afarensis specimens. DeSilva's (2010) argument in favor of a stable midfoot in

PAGE 30

21 early hominins, Au. afarensis included, contradicts Meldrum, et al. 's (2011) interpretation of a mid tarsal break in the Site G prints, which also utilized the reconstructed foot from White and Suwa (1987). Depending on one's interpretation of the morphologies seen in the Au. afarensis pedal specimens, either scenario r egarding midfoot flexibility could be subsumed into the species. A great deal of the debates regarding the locomotion of the Laetoli track makers and Australopithecus can be subsumed into broader discussions of the relationship between form and function of extinct species. This is particularly true in footprints, which are ultimately imprinted by soft tissue, but the functional aspects of the pedal bones play a key role in their creation. The issues of sorting out primitive, adaptively neutral, retentions versus traits that are actively maintained selectively and behaviorally is particularly difficult with early hominins S keletal adaptations in the foot play a key role in bipedal locomotion, but the interplay between morphological adaptation and locomotor function extends up through the entire body, in both hard and soft tissue. Researchers, depending on whether their perspective focuses on selective history or actual functional abilities, can interpret the total body plan in strikingly different ways (Ward 2002). While the morphological form of fossilized remains can provide some clues as to its function, caution needs to be taken, particularly when making inferences on adaptive significance. As Ward (2002 ) notes, there are numerous examples in the fossil record of a disconnect between form and function, par ticularly in primate dentition, although this could easily extend into the postcrania. To what extent any individual morphological differences suggest b ehavioral and functional differences, and if this comprises intraspecific variation or species level distinction is another theoretical and interpretive point of discussion.

PAGE 31

22 Pliocene Hominin Adaptive Radiation and Australopithecine Locomotor Variation While Au. afarensis is the only currently known species at Laetoli contemp oraneous with the footprint tuff Tuttle (1987; 1990 ;1991b; 2008) maintains the possibility of another, yet unknown, species of hominin may be responsible for creating the Site G tra ckway. Locomotor and morphological debates surround ing Au. afarensis aside, our expanding knowledge of Pliocene hominin diversity has made this notion increasingly plausible. Speciation and the amount of accepted intraspecies morphological variability is a continuously debated issue within studies of human evolution. Au. afarensis, as a commonly accepted hypodigm, spans an enormous geographic range across East Africa, and a substantial temporal range of roug hly 7 00,000 years, potentially longer if its anag enetic lineage with Au. anamensis is accepted (Kimbel, et al. 2006). The far reaching and long lived success of this hominin could be interpreted as it being an adaptive generalist, both in body plan and in behavior, with morphological variability within t he taxa explained by sexual dimorphism and temporally progressive evolutionary change (Kimbel, et al. 2009; Ward 2013). However, given the discovery and broadening acceptance of additional contemporaneous hominin species, including Kenyanthropus platyops Au. bahrelghazali and Au. deyiremeda (Brunet, e t al. 1996; Leakey, et al. 2001 ; Haile Selassie, et al. 2015), as well as the rec ently discovered Burtele foot (F ig 2.3), which d iffers markedly from composite feet of Au. afarensis (Haile Selassie, et al. 2 012), it is likely that there was much greater diversity within the hominin clade in the Pliocene than previously interpreted. The interpretation of Au. afarensis as an a daptive generalist with varying morphologies may in fact mask the existence of more th an one species at s ites such as Hadar and Laetoli. The r ec ently discovered Burtele foot ( Haile Selassie, et al. 2012 ) (Fig. 2.6) provides direct evidence of morphological variability among hominins prior to 3.0ma. Dated at 3.4ma, BRT VP 2/73 from the Woranso Mille study area in Ethiopia is contemporaneous to Au. afarensis at nearby Hadar. Consisting of a small associated collection of metatarsals and

PAGE 32

23 phalanges, the foot is directly comparable to specimens such as A L 333 115. Despite sharing certain bipedally adapted features with australopithecines, t here are numerous unique aspects to this specimen. Its potentially adductable hallux is particularly perplexing, given its age The base of Mt I is tall and deeply conc ave, lack s the dorsal doming of the head seen in australopiths, and is short relative to the other metatarsals. This ratio puts Figure 2.6 : BRT VT 2/73 partial foot from Burtele, Ethiopia. From Haile Selassie, et al. (2012). the Mt I of BRT VP 2/73 in the morphological range of extant apes, and as such was likely not used for toe off. The second ray, which consists of the Mt II and proximal and intermediate phalanges, further evidences the abducted nature of the hallux. Its torsion towards the hall ux is less than seen in extant apes, but is significantly more than that of modern humans (Haile Selassie, et al. 2012). The fourth ray is unique in that its metatarsal lack s the expanded stabilizing base morphology seen in later hominins, including Au. a farensis ( Ward, et al. 2012), but also in its longer length than both Mt I and Mt II, which is a unique potentially derived, condition not seen in any extant primates. All of the phalanges of the Burtele foot exhibit strong curvature, although less than in Pan, and generally resemble

PAGE 33

24 those of Ar. ramidus and Au. afarensis (Haile Selassie, et al. 2012). In essence, this specime n appears to continue the pattern seen in Ar. ramidus but with continued refinements for bipedality, and stands in contrast to Au. afarensis. In addition to the Burtele foot and its clear morphological differences from other reconstructed australopitheci ne feet, there is growing evidence, particularly from South African sites, for the morphological variability within the australopithecines in the Plio Pleistocene. While Au. africanus has been argued to continue the same basic locomotor pattern as Au. afar ensis (Ward 2013), the Stw 573 ("Littl e Foot") skeleton, dated at 3.5M a, has been argued by its discoverers to not only constitute another species ( Au. prometheus ) but also to display noticeable morphological differences (Clarke 2013). While most of the joint morphology and orientation does not differ greatly from A. afarensis or even H. habilis the hallux is argued to be slightly divergent, which would suggest abductability. Additionally, the cuboid calcaneal articular surface is bowl shaped, which in addition to the more archaic morphologies of the lateral cuneiforms, points towards greater mid foot mobility. However, the non weight b e aring navicular suggests that Stw 573 had the presence of somewhat of a longitudinal arch (Clarke 2013). Au sediba a lthough geologically much younger than most known australopithecines, dated at 1.98Ma, similarly has a mix of primitive and derived traits in its locomotor complex, including its feet (Zipfel, et al. 2011). The talus exhibits some human like morphologie s, but the neck angle and torsion angle of large talus head are both argued to be more ape like, as is the lateral plantar process, which is positioned more superiorly, not in a weight b e aring position. This would potentially result in elevated heel stress especially in light of its more mobile calcaneocuboid joint, relative to modern humans. Its calcaneal tuber is significantly more gracile than A. afarensis and modern humans, but contains an attachment for a long plantar ligament, which stabilizes the hu man mid foot. The calcaneus overall resembles the ape like morphological condition moreso than that of modern humans. However, the

PAGE 34

25 metatarsals largely resemble humans in morphology, and the positioning of the talus strongly suggests the presence of an arch This odd mix of traits, particularly the positioning of the lateral plantar process, has led some researchers to suggest that A. sediba practiced a form of bipedalism unique from both A. afarensis and modern humans (Zipfel, et. al 2011). A recent study o f calcaneal robusticity amongst australopithecines demonstrated marked differences between certain species, which was argued to demonstrate differences in mobility and habitat exploitation (Prang 2015). The taphonomic bias of the fossil record towards cra niodental remains has led to further bias in studies of early hominin variability and diversity, and the small size of the australopithecine postcranial collection has limited researchers to some extent (Ward 2013). However, if researchers are beginning to accept Pliocene hominin dietary diversity connected to craniodental morphological variation, it is not a stretch for morphological variability connected to the locomotor suite of adaptations to be similarly accounted for in the fossil record (Haile Selass ie, 2016). Ecological context plays a significant role in driving adaptation, particularly as it concerns mobility, diet, and subsistence behavior (Potts 2007). Given the interconnected nature of these aspects, a high degree of environmental variability, on both the regional and local habitat scale, would provide impetus for an adaptive radiation in both cranial and locomotor morphology. The degree of environmental variability in Pliocene Africa is well documented from faunal, ocean core, and geologic ev idence, as is the existence of australopithecines in a wide range of paleohabitats (Behrensmeyer and Reed 2013). The degree to which hominin adaptation to habitat and environmental variability could result in speciation, particularly as it concerns Au. afa rensis, is nevertheless a subject of debate. Behrensmeyer and Reed (2013) argu e that as a genus, australopithecines were remarkably adept at exploiting different environments, and leave open the possibility of multiple hominin species co existing at sites, despite maintaining the interpretation that the specimens from

PAGE 35

26 Hadar, Laetoli, and Woranso Mille are all Au. afarensis. Foley (2002 ) disputes the adaptive radiation hypothesis for australopithecines, and instead posits a dispersal based divergence of ea rly hominin species. However, this is largely based on broad time averaged geographic evidence, as well as the "course grained" nature of the pre Homo species record. While dispersal and patterns of mobility among hominins no doubt influenced adaptation, gaps in the fossil record do not preclude habitat level radiation, which does not need to be "explosive" in nature or require intense specialization to allow for speciation (Foley 2002) Given the relatively novel, evolutionarily speaking, nature of homini ns' locomotor capabilities and the environmental variability of Pliocene east Africa, adaptive experimentation can reasonably be assumed to lead to some degree of regional speciation and admixture (Potts 1998). Additionally, if high mobility can be expecte d of early hominins, which the existence of Au. bahrelghazali in Chad points to, it is reasonable to expect populations to have overlap in their dispersals. As Haile Selassie (2016) points out, modern chimpanzees and gorillas are adaptable to broad variety of niches, and often can overlap in their ranges and exploited environments. At Olduvai Gorge, both robust and gracile hominins existed side by side, and appear to have exploited different ecological niches (Barboni 2014) and it is likely that similar ecological divergence occurred amongst earlier hominins at Pliocene sites, Laetoli included However, the taxonomic validity of many of the newly described P liocene species continues to be a subject o f debate. Much of it boils down to the thorny issue of the use of species concepts within hominin paleontology, and whether or not the variation in the hominin record matches what we presently seen amongst extant primates (particularly apes). As the hominin evolutionary tree becomes bushier and bushier, particularly in the Pliocene, researchers are increasingly having to rectify these issues with the very nature of adaptive radiat ion and the tempo of evolution at both a micro and macro scale (Tattersall 1986; Tattersall 2000; Potts 1998; Foley 2001; Haile Selassie 2016).

PAGE 36

27 Within studies of intraspecies variation, there has been little consensus on the acceptable degree of morphological variability and body size dimorphism in hominins, even early Homo, but it is particularly no table with Au. afarensis. In addition to the early debates and controversy regarding the inclusion of the Laetoli and Hadar collections within a single species (Johanson and Edey 1990), there has been differences of opinion regarding whether or not the Had ar Au. afarensis collection comprises more than one species of hominin, or is simply a high degree of morphological variation and dimorphism in a single evolving lineage ( Harcourt Smith and Aiello 2004). Multiple hominin species occupying the same site co ntemporaneously is well documented within the fossil record, particularly at Olduvai Gorge an d Sterkfontein which has made taxonomic determination of post cranial remains somewhat difficult for researchers (McHenry 1994) The difficulties in parsing out t he early Homo taxonomy, with H. habilis, H. rudolfensis, and H. erectus sensu lato overlapping both temporally and spatially in Pleistocene Africa, is also a continuing source of debate. In the case of Au. afarensis, it is generally conceded that Hadar constitutes a single lineage evolving over the course of ~400ka. If this is the case, then Au. afarensis represents a highly sexually dimorphic species, with a relatively large amount of morphological variation. The sheer amount of hominin fossil material from all stratigraphic layers of the Kada Hadar member, particularly that collected since the mid 1990s return to the site, has been argued to demonstrate the time averaged variation in the lineage, filling in many gaps left by the original round of resear ch (Johanson 2004; Kimbel and Delezene 2009; Ward, et al. 2012). However, there is continued disagreement about the relative level of skeletal dimorphism seen at Hadar. Some authors view the dimorphism as on par with that seen in modern H. sapiens (Reno, et al. 2003; Reno, et al. 2010 ), while other studies have shown an elevated level of dimorphism above modern H. sapiens and Pan ( McHenry 1991; Ruff 2002) or on par, or approaching, that of modern Gorillas (Richmond and Jungers 1995;

PAGE 37

28 Johanson 2004; Plavcan et al. 2005; Grabowski, et al. 2015). Recent discoveries at Woranso Mille, Ethiopia (Ha ile Selassie, et al. 2010 ) could lend further credence to the latter theories, as the partial skeleton uncovered there has a reconstructed stature far exceeding any other known Au. afarensis individual. Outs ide of skeletal size dimorphism some authors have argued that there may potentia lly be more than one form of locomotor repertoire displayed in the Hadar assemblage, pointing towards different species, although it is argued that there is little evidence to support these claims (Harcourt Smith and Aiello 2004). Plavcan, et al. (2005 ) a nd Ackerman and Smith (2007) nevertheless urge caution when approaching research questions of sexual dimorphism and specific variation, as there still exists gaps within the fossil record, and thus our understanding. The issues explored in this chapter a ll come to a head when interpreting the Laetoli Site S hominin prints and the questions that have emerged from their discovery The size differences between the Site S and Site G prints are notable, but are their morphological and gait differences between them? Would any morphological differences suggest functional differences, or simply normal variability in foot construction? Can the Au. afarensis hypodigm subsume an even greater amount of body size and morphological variability, or is their greater homin in diversity at Laetoli than previously assumed? Could the paleoenvironment of Pliocene Laetoli have supported more than one hominin species, and to what extent would they have been divided adaptively and behaviorally? If the Laetoli prints were made by ei ther one or two hominin species, what are the evolutionary and behavioral significances for the population s at Laetoli and early hominins in general? Footprints are immensely useful, but caution needs to be taken when attempting to make further reaching in terpretations based on them. However, when combined with additional fossil data and context, wider patterns can potentially be seen.

PAGE 38

29 CHAPTER III MATERIALS AND METHODS Materials and Data Collection Sample The thirteen hominin footprints from Site S are located within three separate trenches, named TP2, M9, and L8 by the initial impact assessment team, as described in Masao, et al. (2016). The three trenches were re opened under the direction of Charles Musi ba during the summer 2016 field season as part of the UNESCO recommended regular monitoring at the site, in order to assess the trackway and capture the photographic data utilized in this analysis. The footprints themselves are of highly variable preserva tion conditions. Several prints are incomplete, partially faulted or distorted, or filled with dense, difficult to extract sediment, similar to that seen in several of the prints in the G Trackway (White and Suwa 1987). Nevertheless, some of the prints are in better preserved, more informative conditions. The comparative data from the Site G trackways were taken from the first generation cast of the southern portion of the trail housed in the Laetoli visitors center at Locality 8, also during the su mmer 20 16 field season. Later hominin and modern human comparative data was culled from published data from numerous footprint related studies, including measurements from Ilere t, Kenya ( Dingwall et al. 2013), Peru (Tuttle, et al. 1990), and worldwide (Tuttle, et al. 1987). A substantial comparative data set of modern unshod human footprint metrics from the Hadzabe of Tanzania, collected by Dr. Musiba used in the Musiba, et al. (1997) study was given for use in this research The use of data from habitually uns hod or minimally shod modern humans provides a much more directly comparable data set to use with early hominin footprints, as it has been demonstrated that habitual footwear use results in different foot morphology, particularly in foot breadth and relati ve hallucial gap (Tuttle 1990). Metrics included in this study were footprint length, heel

PAGE 39

30 width, ball width, foot angle, step length, stride length, stride width, and stature from 52 adult and juvenile individuals of both sexes Descriptions of Site S Hominin Footprints Until the full extent of the Site S trackway and total number of prints in sequence can be determined, maintaining the naming structure of the individual prints as laid out by Masao, et al. (2016) is best for the sake of consist ency. The prints are designated by the individual track maker (of which there are arguably two) and their step sequence, and qualified by the trench from which they are located. TP2 S1 1 : This is the best preserved print of the three from the TP2 trench, and it has minimal fracturing through the print itself. The posterior portion of this right footprint is cut through by a calcite filled vein running through the heel somewhat mediolaterally, as is the medial an terior portion, although these veins appear to have caused little distortion to the morphology of the print itself. The print is characterized by a markedly anteriorly placed hallucial imprint, as well as a noticeable posterior heel drag mark. There appears to be a small bovid print overlaying the p rint at the anterior portion of the drag mark/posterior portion of the hominin print. TP2 S1 2 : Roots and an east west fault have affected the preservation of this left footprint, and the anterior portion is shifted downwards and morphologically distorte d. Only some basic measurements could be taken as the state of preservation would prevent any detailed morphological or locomotor analyses. TP2 S2 1 : The smallest of the Site S footprints found s o far, it is comparable in size, but generally larger than some of the G 2/ 3 footprints, and is clearly from anoth er individual. It is noticeably distorted, as the anterior portion is remarkably broad. This could be from the individual slipping slightly during gait, or taphonomic factors could be the culprit, as there is a significant amount of roots crossing through the print itself. Masao, et al. (2016) have

PAGE 40

31 listed this as a right print, although the preservation status has made this somewhat indeterminat e. The lack of a corresponding second footprint along th e line of progression lends credence to their identification. However, the uplifting in the trench directly above the print should not be discounted. M9 S1 1 : This left footprint was still largely filled with dense, compacted sediment at the time of data collection, an d the anterior portion is criss crossed with two veins filled with calcite. Basic measurements could still be taken, however, and once the adhering matrix eventually is removed, more detailed analyses can likely be done. M9 S1 2 : This right foot print is remarkably well preserved, despite the calcite filled fractures on the anterior portion. The morphology and shape of the print is well defined in the tuff layer, allowing for some kinesiological inferences. The hallucial impression is noticea bly anteriorly placed, but not as much as in TP2 S1 1, and does not appear to be divergent, although the unfortunate placement of calcite vein may shroud this. M9 S1 3 : Similar to the previous print in the sequence, this left footprint is also well prese rved, although diagonally cut across by a calcite filled vein that unfortunately terminate s and splits through the halluc al imprint. Despite this, the hallux does not seem to project as much as in M9 S1 2. M9 S1 4 : As in M9 S1 1, this print is still fille d with the dense compacted matrix (if not more so). Although if the previous two prints in the sequence are any indication, once this matrix is carefully removed, the print will display the same degree of preservation. There are still two calcite filled ve in s cutting through the print, however, one diagonally from the what is likely the lateral ball of the foot, through the edge of ray IV or V. L8 S1 1 : While some of the L8 prints are decently preserved, they are generally more affected by taphonomic fact ors than those in the other trenches, likely a result of the footprint tuff's proximity to the surface in the L8 trench. L8 S1 1 is a right footprint, heavily eroded and bearing vein distortion Few kinesiological inferences can be made, although

PAGE 41

32 some weig ht distribution characteristics and basic length and width measurements can be taken. L8 S1 2 : This left footprint is criss crossed with vein s, and has some posterior erosion damage, but the shape and morphology of the print is decently preserved. There is some degree of medial slippage/displacement the hallux, but minimal to no noticeable divergence. L8 S1 3 : This right footprint is the best preserved in this sequence. Although it has some vein damage, particularly on the lateral anterior side, the morphology of the interior of the print is much more intact. The medial ball imprint is particularly prominent, and as with the other prints in th is sequence, the hallux shows no divergence, and little protrusion from the rest of the anterior print. L8 S1 4 : This left footprint is much more fractured than the previous two, although the shape of the print has remained intact. There is some sembl ance of toe imprints remaining in th e anterior portion, and there is a marked degree of hallucial slippage medially. It is more prominently than in S1 2, although taphonomic factors may have exacerbated its appearance. L8 S1 6 : Due to extensive faulting and tree root damage, the northern portion of the trench has a precipitous drop in footprint preservation. S1 5 in this sequence is completely obliterated, while S1 6 is only preserved by the posterior heel portion, which is barely distinguishable amongst the cracked ash fall tuff layer. L8 S1 7 : This right print has retained most of its shape, although it is heavily eroded and root damaged, particularly in the anterior portion. Basic measurements could be taken, although there was some estimation with th e anterior section. The footprint tuff itself is heavily fractured in this section, and the sequence presumably ends here.

PAGE 42

33 Figure 3.1: Footprints from TP2 Trench. Clockwise from top left: TP2 S1 1, S1 2, and S2 1.

PAGE 43

34 Figure 3.2: M9 Trench Footprints. Clockwise from top left : M9 S1 1, S1 2, S1 4, S1 3.

PAGE 44

35 Figure 3.3: L8 Trench Footprints. Clockwise from top right: L8 S1 2,S1 3, S1 4.

PAGE 45

36 F igure 3.4: L8 Trench Footprints. Clockwise from top right: L8 S1 1, S1 6, S1 7.

PAGE 46

37 Photogrammetric Methods Use of photogrammetry in prehistoric footprint studies The use of digital photogrammetry in paleontological and archaeological studies has greatly increased in prominence in recent years. Advances in capture and processing technol ogy and software have substantially increased the capabilities of researchers utilizin g photogrammetric methodologies, as well as reducing the overall cost and difficulty of its use. These advances have been a boon especially for s tudies of prehistoric footprint s where understanding both the extent and depth profiles of trackways is particularly crucial for a variety of research questions (Matthews 2008 ; Matthews, et al. 2016 ). The use of manual photogrammetric methods to create contour maps of footprin ts was pioneered by Leakey and Harris (1987) on the Site G hominin trackway, and demonstrated the importance of photogrammetry in providing crucial visual datasets for trackway studies (Bennet and Morse 2014). Digital capture methods both laser scanni ng and digital photogrammetry, allow for greater accuracy in reconstruction in three dimensions of footprints as well as the ability to conduct analyses that would otherwise be difficult or impossible to do on the hard prints themselves and visualize qualit ative characteristics that are harder to discern in the sediments The ability to create numerous high resolution three dimensional and orthographic models allows researchers to better visualize larger scale movement patterns of prehistoric fauna, includin g hominins, that would otherwise be time consuming and difficult to plot accurately. While there are slight differences in the results between optical laser scanning and close range digital photogrammetry, the latter has the added benefits of ease of tran sportation and use and cost effectiveness, as it can be done with a relatively simple setup and a decent high resolution digital camera (Matthews 2008; Bennet, et al. 2013, Bennet and Morse 2014 ; Matthews, et al. 2016 ). This is particularly useful for foot print sites, which are often located in remote areas with harsh conditions that would increase the difficult of using bulky, sensitive technology such as laser scanners.

PAGE 47

38 In addition to the utilitarian and direct data collection benefits of using digital photogrammetry to analyze footprint trackways, it s use is crucial for the preservation and conservation of prehistoric footprint trackways like those at Laetoli. Fossilized footprint trackways, having been imprinted upon an ever changing landscape, are pa rticularly vulnerable to the forces of geology and the environment including erosion, perturbation and uplift. At Laetoli, the situation for the innumerable animal trackways is noticeably dire. Even with the conservation steps taken on the Site G hominin trail, it has nevertheless been affected by all of the above forces, and the integrity of the prints has lessened since their original discovery. While strides were taken to protect the Site G hominin trails, the associated animal trails have been conti nuously threatened from pastoralist herding, erosion, and the intrusion of vegetation into the footprint tuff (Musiba, et al. 2008). Like other fossilized footprint trails, the Laetoli trackways are an immensely important data set that provide insight i nto the faunal community and the paleoecological makeup of an area (Lockley 1998; Musiba, et al. 2008). Since there is no way to prevent the natural forces affecting the trackways, photogrammetry is uniquely suited to help p reserve the footprint record in some way for future researchers. Close range photogrammetry has been used not only to document paleontological and cultural site s for posterity and preserve the data digitally, but can also be used to document the effects of natural and anthropomorphic fo rces on the sites (Bates, et al. 2008; Matthews 2008; Bennet, et al. 2013 ; Matthews, et al. 2016 ). Both the Site S and Site G hominin trails are closely associated with the trackways of other animals, and photogrammetric methods can preserve this relationship, even after erosion and trampling have taken their toll. Photo capture The nature of photogr ammetric data collection necessitates photo capture within well lit settings and from multiple angles, in order to sufficiently capture the details and depth of the subject. In the case of footprint trackways, the subject can be

PAGE 48

39 treated as a miniature land scape. The affects of lighting and shadows on depth/height reconstruction as well as accuracy need to be accounted for, in order capture the best possible photograph. The photos of G Trackway cast were captured by myself and Dr. Musiba with a Pentax K10D camera. The camera was mounted upon a roughly two meter telescopic boom which allowed for consistent photo capture height and horizontal positioning by the handler. The camera was also equipped with a remote shutter with which a second person operated the photo capturing, allowing the camera handler to maintain their positioning and control photo overlap consistency. After reviewing and discarding photos with poor lighting and resolution, well over 200 photos were captured and utilized for the G trackway ca st. The Site S trackway photos were similarly capture d on the telescopically mounted Pentax K10D, although I was unfortunately not present for the re opening of the trenches and subsequent photographing, and other members of the team were responsible for the data capturing. Between the three trenches, roughly 300 images were captured for photogrammetric purposes, with roughly half of those dedicated to the larger L8 trench. In all cases, meter sticks and small scales were placed in order to provide calibra tion for the metric data in the resulting photogrammetric analysis. Table 3.1 Photogrammetric Capture Data Track way Capture Height s (m) Avg. Focal Length (mm) Image Resolution (pix) Pixel Size (mm ) # of points in cloud Avg. Root Mean Square Reprojection Error (pix) Site G Cast 1.0 /2.0 33 /18 3872x2592 0.00619046 21.6mil 1. 24 L8 1.0 /2.0 55/18 3872x2592 0.0061904 6 22.9mil 2. 70 M9 1.0m 55 3872x2592 0.00622821 16.9mil 0.412 TP2 1.0m 24 3872x2592 0.00619046 8.5mil 0.201

PAGE 49

40 Photo processing : Two separate programs, Agisoft Photoscan (v. 1.2.4) and Photomodeler Scanner were utilized to create accurate three dimensional photogrammetric models of the trackways. These photogrammetry programs function by recognizing duplicate features and points across numerous overlapping photographs, which allow for the creation of an extractable XYZ "point cloud" with which to build a three dimensional model from. Both programs use similar processing procedures, but function slightly differently from one anoth er, which allowed for a control check on the different models created by the programs. The functionality and capabilities of the programs also influenced what data was extracted from each of the models. These programs allow for a large degree of automation in photogrammetric procedure, as well as manual tweaking and correction of the resulting point clouds and 3 D models. Since no control point calibration was conducted prior to photo capturing, the combination of automated algorithms and manual correction of the models was incredibly useful. The procedure for processing photos in these programs was as follows: 1. Orientation of inputted photos and estimation of camera positions 2. Alignment of photos 3. Construction of sparse point cloud 4. Manual editing of sparse po int cloud to rem ove any incorrect/outlier points 5. Manual marking of points on the photos and creation of known distance scales 6. Recalibration and alignment of photos based on created scales 7. Creation of dense point cloud surface 8. Further manual editing of dens e point cloud surfa ce for accuracy and unnecessary point reduction 9. Triangulation and creation of three dimensional mesh surface.

PAGE 50

41 Figure 3.5: Digital elevation model of the southern portion of Site G Trackway cast

PAGE 51

42 Figure 3.6: M9 Trench photogrammetric contour plot. Bar across the trench diagonally is a two meter stick.

PAGE 52

43 Figure 3.7: TP2 Trench photogrammetric contour plot with digital elevation color map. The tan rectangle denotes the scale used in the model.

PAGE 53

44 Figure 3.8: L8 Trench photogrammetric orthophoto showing direct textures from photos. Note the abundance of faunal tracks also present.

PAGE 54

45 Numerous iterations and edits of the point cloud meshes were done to achieve the most accurate and usable three dimen sional models of the trackways. Longer processing times and frequent technical difficulties arose from Photomodeler which could potentially be the result of available computing power, or the na ture of the data set itself. Additionally, Agisoft Photoscan consistently produced more accurate models that required less manual editing and re processing. As such, practically all data utilized in this study was extracted from the processing conducted i n Agisoft Photoscan, with the exception being the Site G trackway cast Photomodeler largely functioned as an accuracy check, as its recognition of problematic images, typically from resolution or poor overlap, was often more augmented than Agisoft Photosc an. Upon the recognition of missing points in the cloud, or poorly aligned pictures in Photomodeler they could be corrected quickly in Photoscan Post processing and measurement data extraction : Following the creation of the dense point cloud and mesh wit hin Agisoft Photoscan, two additional steps were taken to allow for metric data extraction. The XYZ point cloud coordinates for each entire trackway were exported into ArcMap (v. 10.5), where a color coded elevation maps and contour interval overlays were created. This allowed for heighted visibility of both the hominin and associated animal trackways, and the subtleties contained within the footprint tuff, as well allow visualization of relative depth and weight transfer within the hominin prints. While Ar cMap allows for a great deal of resolution, it is better suited for broad, overarching visualizations and measurements of the trackways. Individual contour elevation maps were created for each hominin footprint by extracting their limited XYZ coordinates a nd uploading them into FootProcessor, an open source program designed by the Bournemouth University team who worked on the Happisburgh footprint trackway specifically for visualizing and measuring footprint data. In addition to individual contour maps, la ndmarks on the footprints themselves could be more accurately located and measured, and the XYZ data for the landmarks were extracted.

PAGE 55

46 Comparative Analyses Footp rint and Trackway measurements Standard measurements of the hominin footp rints were taken following the methods used by Bennet and Morse (2014) (Fig. 3.9 ). For consistency, all measurements were taken from the rim of the footprint. Footprint length was defined as the length from the pterion to the maximum extent the second dig it. Hallucial length can vary significantly even in modern human populations, as well as in an individuals' footprints themselves, based on substrate mechanics, speed, gait, etc thus using the second digit as the terminus of the forefoot can provide more consistency in measurement Maximum heel width and ball width were both taken from the midline (typically the greatest exte nt of their respective imprints perpendicular to the longitudinal axis defined the foot print length ) Instep width was calculated a s the distance between the deepest point of sole pressure and the external edge of the midfoot at its minimal breadth. Step length and stride length were measured as the distance from heel to heel, and stride width was calculated using the same landmarks. Angle of gait was measured in ArcMap by marking the line of progression and the footprint length, utilizing right triangle plane geometry to calculate the angle. Foot index is calculated as the maximum ball width divided by the footprint length.

PAGE 56

47 Figure 3.9: Measurement schemes utilized. Images from Bennet and Morse (2014). Given the overlapping nature of the G 2/3 footprints, only a portion of the G 3 prints were defined enough to allow for more accurate measurements to be conducted, other tha n stride/step lengths. Even then, it is noted that the deformation caused by the overlapping nature may in fact bias certain measurements, such as heel and ball width. Given the overall similarity, size notwithstanding, of the G 1 and G 2/3 prints, this bi as would arguably not affect the comparative analyses significantly. Additionally, even with photogrammetric imaging, it is difficult to parse out with any certainty accurate measurements of the G 2 individual so only measurements of G 3 were utilized Metric Comparison w ith G Trackway and Modern Humans In addition to a visual qualitative comparison between the photogrammetric plots of the Site S and Site G footprints, the combined metric data was transferred into Past (v. 3.0),

PAGE 57

48 a statistical program d esigned for use with paleontological data. The footprint length and estimated stature of the S 1 and S 2 individuals were plotted alongside published data on modern humans, the Site G tracks and the Ileret trackways (Tuttle 1987; Tuttle 1990; Musiba, et al. 1997 Dingwall et al. 2013; Bennet, et al. 2016). Two multivariate analyses were conducted utilizing the Site S, Site G, and Hadza gait study data. A principal components analysis (PCA) was conducted on the sample to determine the variables and combinations of variables most responsible for the variation in the sample. Additionally a multivariate analysis of variance test (MANOVA) was conducted on the metric data including the PCA variables to determine how statistically significant any variation seen between the groups were. A discriminant function analysis (DFA) was then done to test the placement of the S 1 individual a gainst the G track makers based on the metric and principal component variables. S 1, G 1/G 3, and the modern humans were placed in their own separate groups If S 1 displays closer affinities to either the G 1/3 or the modern human data, it could provide insight into differences in foot morphology, gait, and function. Procrustes Landmark/Shape Analysis While the differential preservation of the S 1 footprints precludes certain kinesiological metrics, particularly in three dimensions, a comparison of the overall shape and location of crucial landmarks between the Site S and Site G footprints is capable of being conducted. Given that foot function influences the morphology of created footprints, comparing the morphological differences between the S 1 print s and G 1/G -3 can potentially be functionally illuminating. Following similar procedures laid out in Berge, et al. (2006), 13 homologous landmark points (Table 3.2 ) were placed on the footprint contour maps within FootProcessor, which in addition to provi ding inter landmark distances, also allows for extraction of XYZ coordinates for the placed landmarks. Landmarks were chosen based on their importance in footprint morphological features, as well as their visibility within

PAGE 58

49 the S 1 sample, many of which are less defined than within the Site G footprints. The six best preserved prints of the thirteen Site S tracks were chosen, TP2 S1 1, M9 S1 2, M9 S1 3, L8 S1 2, L8 S1 3, and L8 S1 4 (Fig 3.7) The other tracks were excluded for poor preservation (TP2 S1 2, L8 S1 1, L8 S1 6, L8 S1 7), or their inability to be examined fully due to adhering matrix (M9 S1 1, M9 S1 4). TP2 S2 1 had to be excluded from the analysis because of its marked deformation, preventing accurate placement of several key landmarks. Ten prin ts in total were chosen for the G Trackway sample, five from each set of prints, G1 29, G1 28 G1 27, G1 33, G1 36, G3 18, G3 24, G3 26, G3 27, and G3 28 (Fig. 3.8) Table 3.2 : Landmark placement Landmark Location L1 Posterior Heel Edge L2 Anterior Toe II L3 Medial Heel Edge L4 Lateral Heel Edge L5 Medial Ball Edge L6 Lateral Ball Edge L7 Anterior Hallux L8 Mid sole L9 Lateral sole L10 Lateral edge of toes L11 Heel center of pressure L12 Mt I impression L13 Medial Sole

PAGE 59

50 In order for the pr ints to be compared without the bias of relative size influencing the analysis, the landmark coordinates were converted in Past into Procrustes coordinates. The process of Procrustes fitting translates, rotates, and scales the shape coordinates, superimpos ing them, and thus making them directly comparable. Three different tests were conducted with the Procrustes fitted coordinates. The first was a point kernel density procedure to visualize the relative placement of the landmarks for all prints in the sampl e, allowing for direct visual comparison of the mean shape of each set (S 1, G 1, G 3, G combined). The procedure creates a heat map showing the most frequent areas of landmark placement within and between samples, creating a basic mean outline shape for t he footprints. The second was a two dimensional relative warp principal component analysis and thin plate spline deformation, which examines the variance seen between the sample as a function of mean shape. The relative deformations between the samples poi nt towards the landmarks of the foot most responsible for the differences between the samples. Given the small nature of the sample, it is likely the differences between the Site S and Site G footprints are subtle. Following the creation of the new variabl es from the PCA, a MANOVA test was conducted to determine if the variability between the S 1 and G mean footprints w as statistically significant.

PAGE 60

51 Figure 3.10 : S 1 Prints Utilized in the 2 D Landmark Analysis. Clockwise from top right: L8 S1 2, L8 S1 3, L8 S1 4, TP2 S 1, M9 S1 2, M9 S1 3

PAGE 61

52 Figure 3.11 : Sample of Site G Tracks used in 2 D Landmark Analysis. Clockwise from top right: G1 28, G1 29, G1 33, G2/3 18, G2/3 24, G2/3 27.

PAGE 62

53 CHAPTER IV RESULTS Qualitative and metric comparison with other hominins and modern humans It is obvious upon first glance that the S 1 track maker is much bigger than the individuals responsible for the Site G trackway The S 2 individual's print is also noticeably larger than G 1 or G 3 although possibly close to the highest estimates for the prints of G 2. The S 1 prints average 251mm in length ( Table 4. 1 ) ranging roughly 10mm across the twelve measurable prints in the three trenches, compared to a 180mm average foot length for G 1 a nd 209mm average for G 3. This situates them with the upper reaches of adult Hadza males while the Site G individuals are roughly equal to the Hadza juveniles (Fig. 4. 1 ) Their size also translates into total breadth, with an average maximum heel width of 72.4mm and an average maximum ball width of 97.2 mm The footprints are comparable in size to the larger individuals from the Ileret, Kenya trackways, and can be situated comfortably within sizes of modern human male averages across the globe, including th at of the habitually unshod Hadza (Fig 4. 3 ) Table 4.1: Mean Metric Results Sample Foot length (mm) Max Ball Width (mm) Max Heel Width (mm) Step Length (cm) Stride Length (cm) Stride Width (cm) Foot Angle Foot Index Stature (cm) S 1 251.8 97.2 72.4 57.9 115.2 75.3 6 38.7 167.3 (est.) S 2 225 N/A 71 N/A N/A N/A N/A N/A 150.6 (est.) G 1 180 78.9 65.8 41.6 82.9 68 18.6 43.8 133.6 (est.) G 3 209 85.5 75.9 43.3 87.6 106 4.3 41.5 148 (est.) Hadza ( AD) 237 88.8 58.8 66.4 131.3 61.3 6.2 37.3 155.2 Hadza (JV) 212.2 78.8 50.9 59.2 118.9 61.9 6.6 37 135.1

PAGE 63

54 Figure 4.1: Box plots comparing footprint length of the Hadza with the Laetoli individuals. The left plot shows the distribution based on sex, while the right shows it based on age. In both average step length (57.9cm) and average stride length (115.2cm) the Site S individual also exceeds those of G 1 or G 3, and is comparable to the low end of variation for adult Hadza, but in the upper reaches of variation for juvenile Hadza. The Site G individuals, however, are at the very bottom reaches of Hadzabe variati on, below the averages of even the juvenile Hadza. (Fig 4.2) The average stride width of 75.3mm e xceeded that of the G 1 individual (68mm) and the Hadza (61.9mm), but is less than that seen in the G 3 prints (106mm). However, a ll of the early hominin stri de widths fit within the range of variation seen in the Hadza individuals who participated in the study which varied widely from individual to individual With an average 6 angle of gait, the S 1 individual walked more out toed than G 3 (average of 4.3 in toeing) but significantly less out toed than the G 1 individual (average 18.6 and is comparable to average angle of gait seen in the Hadza sample, who seem to generally walk with some degree of out toeing.

PAGE 64

55 Figure 4.2: Box plots comparing step l ength of the Hadza with the Laetoli individuals. The left plot shows the distribution based on sex, while the right shows it based on age. The Site S footprints are remarkably similar in their visible qualitative functional qualities to the Site G prints but with some notable differences They possess a deep, oval shaped heel strike impression followed by movement along the lateral side of the sole, with transfer of pressure medially a long the metatarsals. There are variable impression depths alongside o f the medial sole, but clear evidence of a medial longitudinal arch with the deepest impression of the midfoot occurring on the lateral edge. As seen in many of the G 1 prints, the deepest part of the metatarsal impressions in several of the prints occurs on the lateral side, and the lateral toe imprints are noticeably shallow although more prominent in S 1 Unlike the G prints, the hallucial imprint is less defined, with the exception of TP2 S1 1 and M9 S1 2, although this may be the result of taphonomic factors, particularly in the partially fractured L8 trench prints. However, the hallucial imprint is noticeably deeper. Compared to the G 1 prints, the hallux does not appear to diverge as much from the second

PAGE 65

56 digit, TP2 S1 1 and possibly M9 S1 3 exclude d. When the stature estimates and foot length of the Site S individuals are plotted along side those of the Site G prints, Ileret footprints, and global modern human data from Tuttle (1987) (Fig. 4. 3 ) the S 1 individual is situated well within the distr ibution of modern human males and the Ileret track maker s while the S 2 individual is in the upper mid range for modern human females, and near the uppermost estimate for G 3 track maker (Bennet, et al. 2016). Even at the lowest stature and size estimate for the S 1 individual, it is clear that they were likely significantly larger than either of the Site G track makers. Figure 4.3 : Stature and foot length plot comparing hominin individuals with global data of modern humans. The principal components analysis (PCA) showed that two principal components were largely responsible for the variation in the sample, the remainder evenly distributed th rough an additional seven principal components. The variables utilized in the PCA were stature, footprint length, heel width, ball width, step/stride length, and foot index. Foot angle

PAGE 66

57 and stride width were excluded from the final analysis, as their withi n group and individual variability was creating noticeable skewing of the results. PC1, with an Eigenvalue of 656.2 explained 70.3 % of the variance, while PC 2, with an Eigenvalue of 178.3 explained 19.09 % of the variance. Based on the loadings, PC1 large ly represents on footprint length and gait while PC2 largely represents the expected gait given the overall sizes in the sample The PCA plot (Fig. 4. 4 ) shows a clear separation between the G, and S 1 samples, while the Hadza maintain an e ven, but cluster ed distribution across the center of the plot, which is to be expected considering the mix of men, women and juveniles in t he sample. The S 1 tracks show a much closer affinity to the Hadza sample than to the Site G individuals Figure 4.4: PCA Plot with 95% ellipses showing all footprints used in the sample The Discriminant Function Analysis (DFA) (Fig. 4. 5 ) maintains this affinity, demonstrating the degree to which S 1 clusters with the Hadza sample The first function, which corresponds with PC1, displayed an Eigenvalue of 1.673, while the second function, corresponding to PC2, displayed an Eigenvalue of .574. The Wilks Lambda, which was

PAGE 67

58 below 1, showed the separation between the three groups along these two axes was statistically significant, but not highly discriminatory ( =.238, p=<.005, df=4). However, the relatively high canonical correlation for both functions (r=.791 and r= .604 respectively) demonstrates that there is a strong association bet ween the discriminant scores and the groupings. The DFA largely placed the three groups within their own classifications, particularly the Site G group, which remained entirely in its original classification. The Hadzabe sample had several prints classifie d within the S 1 group, which, given their degree of overlap, is to be expected. Several Hadza prints were classified within G, either from juveniles or a single adult outlier. Figure 4.5: DFA plot with 95% ellipses Note the adult Hadza outlier's affin ity with Site G. The MANOVA of the sample based on the first two principal components confirms the statistical significance of the variance within the groups (p=<.005, =.763, F (2, 86) = 13.357). Interestingly, when the three groups are compared agains t each other in posthoc tests, there are different significances based on the two principal components. With PC1, both S1 and the Hadza samples are significantly different from the G sample (p=<.005), but

PAGE 68

59 not from each other (p=.530). However in PC2, the d ifferences between all three groups is statistically significant (p=.001) Kernel Density and Procrustes Landmark Analysis and PCA The kernel density plot of the 13 landmark point shows clear differences in the breadth, particularly in the midfoot and for ef oot region, of the S 1 (Fig. 4.7 ) and Site G (Fig.4.8 ) footprints, even when controlling for overall size. The below density plots show the areas of greatest concentration for landmark placement, visualized as a heat map. Additionally the G 1 and G 3 pri nts show a greater degree of divergence between the first and second digits than in S 1, as well as a somewhat less anteriorly placed center of heel pressure. The results of the two dimensional PCA of the relative warp between the two samples show a much m ore even distribution of variance than seen in the metric data PCA, with greater overlap between the two mean footprint shapes (Fig. 4.6 ) Fig. 4.6 Plotting PC1 Vs. PC2

PAGE 69

60 Fig. 4.6: S 1 Pri nt Landmark Kernel Density Plot Fig. 4.7: G 1/G 3 Avg. Landmark Kernel Density Plot

PAGE 70

61 Figure 4.9: 2 D PCA Relative Deformations. Clockwise fro m top right: PC1, PC2, PC3, PC4 PC 1 (Fig.4.9) with an Eigenvalue of .0026 explained 26.8% of the variance, largely focuse s on midfoot breadth and halluc al gap, based on the relative deformation plot. PC 2 (Fig. 4.9), with an Eigenvalue of .00194, explaining 20.02% of the variance, PC3 (Fig 4.9.), with an Eigenvalue of .0013, explaining 13.6% of the variance, both seem to center around overall breadth of the foot from different landmark distances. PC4 (Fig. 4.9), with an Eigenvalue of .0011, explaining 11.8% of the variance, seems to center exclusively on instep width.

PAGE 71

62 Figure 4.10: Overall Deformation of Mea n Shapes from S 1 Avg. to G Avg The overall thin pla te spline deformation plot (Fig. 4.10) confirms that halluc al gap and midfoot breadth are the primary differentiating variables in the mean shapes of the S 1 and G 1/G 3 individual The low Eigenvalues and relatively even distribution of variance a cross multiple principal components confirms that the differences between the mean footprint shapes are subtle. The MANOVA on the PCA results show the differences are statistically significant, but with weakly associated variation, with the Wilks' lambda t est showing a p value of .002 ( =.336, F (4, 15)=7.426). Summary of Results: The combined PCA and DFA tests show that there are significant morphological differences separating the S 1 and G footprints. While there are similarities in foot pressure distrib ution, and in stride/step lengths, in overall metrics and principal components, the S 1 individual aligns more closely with the Hadza s ample than with the Site G indi viduals. In mean footprint shape, the S 1 individuals show a much narrower foot, particula rly in the midfoot, and a less diverged hallux. These differences are significant enough to discr iminate between the two samples.

PAGE 72

63 CHAPTER V DISCUSSION The results of the comparative print analyses are incredibly intriguing in their implications. While the sheer size of the prints, as well as the reconstructed stature of the S 1 individuals demonstrate that this was clearly a different, noticeably large r, individual than the Site G track makers, there is also potentially differences in foot morphology and l ocomotion. The Site S sample, while overlapping with both some of the Hadza sample and with the larger G 3 prints (and theoretically the G 2 prints), consistently plots in between the two samples in both the PCA and DFA largely skewing closer to the Hadz a sample especially in the DFA The PCA plot shows the three samples distributing along the axes based on foot size and expected mobility. The Hadza sample shows a fairly even linear distribution along the center of the intersection of the two axes, with the juveniles, possessing smaller feet and the expected shorter strides to the left, and the adults with larger feet and larger strides to the right of the plot. The G 1 and G 3 individuals both cluster on the left side of the plot, reflecting their small er size and smaller strides. The G 1 and G 3 prints form their own individual clusters based on the size differ ences between the individuals. Whereas based on their size, the G 1 prints can be expected to shorter strides, albeit ones shorter than a modern human of equivalent size, the G 3 prints have similar relative stride lengths, despite being made by a larger individual. The S 1 print cluster in a similar position on the PC2 axis as G 3, but plot further right on the PC1 axis, partially aligning with t he Hadza sample Their location on the upper end of the PC2 axis indicates that despite their large size, they nevertheless have smaller strides than would be expected for individuals of their size. The discriminant function analysis plot based on the fi rst two principal components (which accounted for 87.4% of the total variance) affirms these differences. The prints of the

PAGE 73

64 two G individuals cluster closer together along the left side of first axis, maintaining their distinctness from the S 1 and Hadza clusters. The S 1 cluster displays even greater affinity towards the Hadza sample, which is affirmed by the statistical analysis showing that these two samples are significantly more different from the Site G sample than they are from each other, at least within PC1. The predicted group membership results seems to have di fferentially discriminated the outlier prints for their closer affinities with the further ends of each groups principal component variation. Th e qualitative comparison shows similar foot function in terms of footprint formation, and the preservation of the currently excavated Site S prints would necessitate a very cautious approach to any pressure depth analysis hence why only a 2 D landmark analysis was conducted in this study. However, the comparisons of the mean shapes of the footprints based on the 13 chosen landmarks demonstrate that there are subtle, yet significa nt differences in the footprint morphology between S 1 and the Site G individuals. The principal components analysis sho wed a fairly even distribution of variance across several principal components, although largely within the first four. The 2 D deformation of the principal components confirms the even distribution of variance across the whole footprint, but shows consist ent differences in footprint breadth and hallucial gap. The primary point of breadth deformation is located in the medial midfoot, which would correspond with the location of the medial longitudinal arch. These significant differences between S 1 and the G individuals cannot solely be accounted for by overall size. The S 1 prints in mean shape are still relatively broad, but more closely resemble the more narrowed shape of modern human footprints as described in Berge, et al. (2006). The S 1 prints show add itional affinities in their phalangeal positioning. The hallux is more closely aligned with the second digit, and the lateral toes decrease in size sequentially from the hallux, where as in the G prints, the lateral toe imprints are as long, and in som e ca ses, longer than the halluc al imprint. The differences in footprint shape point towards differences in foot morphology

PAGE 74

65 between the individuals. W hile the differences in halluca l gap have more obvious physiological correlates, the mid foot differences lik ely reflect arch development and construction, in whi ch the S 1 individual possessed a higher, more human like medial longitudinal arch. While qualitatively the distribution of pressure in all three sets of Laetoli prints ap pear to be remarkably similar, without an quantitative analysis of pressure distribution, it is difficult to make detailed inferences about any kinematic differences between the S and G individuals. Nevertheless, in overall shap e, the S 1 footprints reflect an arguably much more huma n l ike construction to the foot, and it would not be outlandish to suggest that some functional differences would be expected. The results of the discriminant function analysis, which plotted the S 1 individuals away from both the G individuals and nearly wi thin the Hadza sample, suggests that the S 1 individual differed in locomotor pattern from both groups although the latter less so The issue at hand then, is what these differences mean as far as taxonomic variation in the Pliocene Laetoli hominins. Mo dern humans are not without great variation in foot morphology, functi on, and footprint shape. Halluca l gap, arch development, toe length, and foot breadth are all dependent on a variety of factors, both physiological and environmental. Habitually unshod populations consistently display broader feet and a larger halluc al gap than groups that consistently wear shoes, and while it is sometimes argued that they display lower arches, studies such as Tuttle, et al. (1990) and Musiba, et al. (1997) demonstrate t his is not necessarily the case. There is also variation between unshod populations, as Musiba, et al. (1997) demonstrated with the Hadza be which on average have longer an d narrower feet than the Machiguenga people studied by Tuttle, et al. (1990). Musiba et al. (1997) also remark that the Hadza, for their stature, display longer strides and quicker gaits than seen in comparable populations. However, it is important to note that while there are some variations within populations, mostly age related, the mean of many metrics and shape wise comparis ons are largely similar between individuals of both sexes. In the comparative PCA,

PAGE 75

66 Hadza men and women almost entirely overlapped with one another, and while juveniles skewed closer to the Site G sample, the latt er still was largely out of the Hadza range of variation. Within the G prints, the variability between the prints and between the track makers is fully within the expected v ariation seen in a population, based on the Hadza data. Despite the size differenc e between G 1 and G 3, they maintai n the same mean footprint shape Conceivably the G 2 print would fall within the same mean shape were it better visible While the S 1 individual displays some affinities with the G individuals, consideration needs to be made of the significant differences in footprint shape. While subtle but significant differences in footprint morphology are on the surface not incredibly differentiating, when contextualized with their size and the known paleontological record, a patter n can be glimpsed. It is abundantly clear that the S 1 individual is a surprisingly large hominin for the height/foot length ratio (Tuttle 1987), the Hadza sta ture based regression (Musiba, et al. 1997), or primitive body proportion based (Dingwall, et al. 2013), S 1 is significantly larger than either G 1, G 2, or G 3, while S 2 is larger than G 1 or G 3, and likely G 2. The stature and presumptive size of S 1 greatly exceeds that of any known Au. afarensis individuals from Laetoli or Hadar. The largest known individual assigned to the Au. afarensis hypodigm is the partial skeleton from Woranso Mille, KSD VP 1/1 ("Kadanuumuu"), which is also notable for its si ze, described as being the size of a small modern human (Haile Selassie, et al. 2016). Masao, et al. (2016) give an estimated stature of 158cm for "Kadanuumuu," likely based off the tibial length, but do not provide their method of stature estimation, as neither Haile Selassie, et al. (2010) nor Lovejoy, et al. (2016) give a direct estimation of the stature of the individual. The implication that the Woranso Mille individual's lower limbs were elongated, despite having an Au. afarensis like upper body plan (Haile Selassie, et al. 2016) may influence what the reconstructed height of this individual would be Additionally,

PAGE 76

67 while "Kadanuumuu" has been assigned to Au. afarensis based o n shared morphological traits, no associ ated craniodental remains have pub lished for the individual, which prevents definitive taxonomic assignment. However, even if one accepts this estimation, S 1 is still taller at the lowest end of its estimated stature. There are othe r important implications of the size and morphology of the S 1 prints. Alongside "Kadanuumuu," the Site S hominin prints demonstrate that there were large bodied, bipedally advanced hominins in the Pliocene. While previously "Kadanuumuu" could be constr ued as an exceptionally tall outlier, much as "Lucy's" diminutive stature has been, the S 1 prints confirm that this was not the case. Coupled with the existence of small bodied hominins such as H. naledi, H. floresiensis, and Au. sediba in the Pleistocene the notion of a linear trend of increasing body size and postcranial change in hominin evolution is becoming increasingly harder to defend. Grabowski, et al. (2015) and McHenry and Brown (2008) also both posit this from different morphological perspectiv es, which stresses the point that hominin evolution, particularly the transition from the australopithecine grade to Homo, was not a tidy evolutionary affair. One does not need to believe in a much more speciose hominin evolutionary tree to recognize that adaptive responses of the Plio Pleistocene hominins was likely not driven by unidirectional forces. However, this leads to anothe r important, more controversial, implication of the Site S prints: the taxonomic designation of the track makers. Given their contemporaneity with the G prints, and the latter's widely accepted attribution to Au. afarensis it is not an unreasonable assumption that the Site S prints were also made by Au. afarensis, albeit a much larger one. While the Hadar Au. afarensis assembl age has been argued to display a noticeable amount of s exual dimorphism in body s ize (Kimbel and Delezene 2009), the differences within that sample are dwarfed by the estimated differences between S 1 and G 3. Presuming that S 1 was a male, and the Site G indi viduals were females and a juvenile as Masao, et al. (2016) have interpreted, this is an enormous amount of sexual dimorphism

PAGE 77

68 within a species. This level of dimorphism is not unheard of in primates, as both gorillas and orangutans have significant si ze differences between males and females. This alone would have significant evolutionary and possibly behavioral implications as far as social structure is concerned, which Masao, et al. (2016) subsume into their interpretation of the track site. The noted differences in footprint morphology, which suggest differences in foot structure, c ould suggest locomotor dimorphism, wherein the larger males were more advanced bipedally than female individuals, a notion entertained by some researchers of Au. afarensis (Harcourt Smith and Aiello 2004). Considering the significant differences in the Laetoli track sets, and the lack of Au. afarensis fossils other than "Kadanuumuu" that approach the reconstructed size of S 1, it is imperative to consider alternate interpre tations regarding the track makers at Site S. Given the emerging picture of hominin diversity in the middle Pliocene, both in terms of speciation and postcranial variation, the likelihood of another hominin species inhabiting Laetoli is not an outlandish proposition. Laetoli is already noted for its relative rarity of hominin remains, Au. afarensis included (Su and Harrison 2008), and the taphonomic and environmental causes for their rarity may mask taxonomic diversity. The animal trackways have provided crucial paleoecological evidence fleshing out the faunal community of Pliocene Laetoli, particularly for species that may not be fully represented by the fossil record (Musiba, et al. 2008), and it is logical that the same would be true for hominins as wel l. While the Ileret footprints are commonly attributed to H. erectus their discoverers note that H. habilis and Paranthropus boisei are also known to have exist ed in the area contemporaneously with the footprints, and are just as likely to have been the track makers (Dingwall, et al. 2013). The existence of K. platyops and other australopithecines in East Africa in the middle Pliocene roughly contemporaneously with Au. afarensis points towards potential overlap in range and habitat among species. The disc overy of the first fossils of P. aethiopicus outside of Ethiopia and Kenya in the Ndolanya Beds at Laetoli is also an

PAGE 78

69 important development to be considered (Harrison 2011). While known estimated ages for P. aethiopicus, the oldest k nown robust australopit hecine, range from ~2.3 to 2.7 Ma, the fossils from Laetoli are the oldest securely dated specimens at 2.66 Ma (Harrison 2011). The earliest known fossil evidence does not necessarily correlate with its first appearance date, and the emergence of the robus t australopithecines is still phylogenetically murky. The possibility of earlier hominins with affinities to P. aethiopicus existing at Laetoli contemporaneously with Au. afarensis is entirely plausible. Paleoecological interpretations of Laetoli during the mid Pliocene also support the notion of potential overlapping hominin populations At Olduvai Gorge during the Pleistocene, H. habilis and P. boisei are known to have occupied the area contemporaneously, and are argued to have occup ied different ecological niches This appears to be based on dietary and habitat preferences, as isotopic analyses on dental enamel demonstrate that P. boisei 's diet consisted of a substantially higher proportion of C4 grasses and sedges than H. habilis (Barboni 2014). Andrews and Bamford (2008 ) note the frequent association of P. aethiopicus at Laetoli with habitats dominated by C4 plants Given an accepted interpretation of a dynamic, mosaic like paleolandscape that included the gradual expansion and periodic dominan ce of C4 grasses at Laetoli, it is possible that hominin adaptive responses to thes e factors resulted in speciation Alternatively, the periodic disruption of woodland environments could have resulted in temporary range and habitat expansion for hominin sp ecies more adapted to more open grassland environments and diet Tuttle et al. (1990) have disputed Au. afarensis as the Site G track maker because of its pedal morphology, and posited the existence of another, more bipedally advanced hominin residing at Laetoli. Given the more advanced appearance of the S 1 prints, their closer statistical affinity with the Hadza than with the Site G individuals, and the difficulty in rectifying their size within the known size distribution of Au. afarensis this notion is worth revisiting.

PAGE 79

70 As it stands, there are three possible scenarios for the taxonomic d esignation of the Laetoli print makers The first, p osited by Masao, et al. (2016), is that both the Site S and Site G pr ints were left behind by Au. afarensis with S 1 representing a large male individual, and S 2 and the Site G prints were likely left behind by female and/or juvenile individuals. This would vastly increase the known body size distribution of Au. afarensis and suggest a g orilla like sexual dimorphis m within the species. The differences between the two sets of prints are either normal distribution influenced by allometry, or indicative of locomotor dimorphism between the sexes, with males possessing more derived bipedal capabilities. S 1 could also p otentially be a peculiar individual with much larger feet than expected. The second scenario is that another, larger, hominin species existed alongside Au. afarensis at Laetoli, and is responsible for the Site S prints. This species would potentially bear some resemblance to Au. afarensis in locomotor pattern, but display different pedal morphology, reflected in the more human like shape and size of the S 1 prints. These individuals could potentially be an early paranthropine, K. platyops or another gracil e species of Australopithecus A third equally controversial, potential scenario, is that neither Site S nor Site G trackways were created by Au. afarensis Based on the arguments provided by Tuttle, et al. (1990) and other researchers that the species was unlikely to h ave created the prints, another, more bipedally derived hominin species was responsible for both trackways. Like the first scenario, the differences between the track morphologies are subsumed in normal variation or are a function of allo metry. While all three scenarios are likely, this third scenario would be a much more radical shift in the interpretation of the trackways. While attribution of the Site S prints to Au. afarensis is a valid interpretation, based on the significant differen ces in size and morphology of these prints with the G trackway, it is equally parsimonious to argue for the presence of another, larger hominin species at Laetoli that could be responsible.

PAGE 80

71 CHAPTER VI CONCLUSION The serendipitous discovery of the Site S p rints is tremendously exciting. Despite being one of the most important sites for understanding hominin evolution and paleoecology and researched for decades, Laetoli is yielding surprises even within a stone's throw of the Site G trackway, its most promi nent contribution. Given the size of the site and the extent of the footprint tuff, additional hominin prints are no doubt waiting to be uncovered alongside the innumerable animal footprints blanketing many localities within Laetoli. The Site S prints have considerable implications for hominin evolution in the Pliocene, as well as the evolution of bipedality within the hominin clade. While the interpretations of their evolutionary, locomotor, and behavioral significance are open for debate, the ir presence a rguably confirms the existence of large bodied hominins in the Pliocene fossil record, much earlier than previously believed. The research undertaken in this thesis has demonstrated several important results. In addition to placing the foot size and reco nstructed stature of the Site S 1 individual well into the range of modern human males the principal component analysis and discriminatory function analysis conducted demonstrate the statistical placement of S 1 in between the Site G individuals and the m odern human Hadza sample, with greater affinity towards the latter. The MANOVA demonstrated that the differences between these samples ar e statistically significant, particularly between S 1 and the Site G individuals. While qualitatively, the sets of foot prints appear similar in form and function, the Procrustes landmark PCA showed significant differences between the Site S and Site G footprint shapes. S 1 displays greater hallucial abduction, phalangeal lengths that decrease laterally from the hallux an d a narrower midfoot. Overall the shape of S 1 appears more derived towards the modern human condition than G 1 or G 3. The differences are subtle, but are shown to be

PAGE 81

72 statistically significant. While these could be subsumed into normal variation, it is al so potentially indicative of differential foot construction and locomotor function. When the results of this research are viewed cumulatively, and put in the paleoecological context of Pliocene Laetoli, the implications become much more complex. While si ze and differences in footprint shape alone cannot definitively demonstrate taxonomic differences, this rese arch provides evidence that it is entirely plausible that another, larger species of hominin co existed with Au. afarensis at Laetoli, and is responsible for the creation of the Site S prints. Given our current knowledge of the Pliocene fossil record and paleoecological drivers of variation and adaptation, an attribution of the Site S prints to Au. afarensis without consideri ng alternative hypotheses is somewhat premature. Further analyses of the existing prints, as well as the uncovering of additional prints and recovery of further hominin fossils will hopefully elucidate the emerging picture of hominin locomotor diversity at Laetoli.

PAGE 82

73 REFERENCES Aiello, L. and Dean. C. 1990. Human Evolutionary Anatomy. Academic Press Limited Ackermann, R.R., Smith, R.J. 2007. The Macroevolution of our Ancient Lineage: What We Know (or Think We Know) about Early Hominin Diversity. Evolu tionary Biology 34: 72 85. Andrews, P. and Bamford, M. 2006. Past and present vegetation ecology of Laetoli. Journal of Human Evolution 54, 78 98. Barboni, D. 2014. Vegetation of Northern Tanzania during the Plio Pleistocene: A synthesis of the pale obotanical evidences from Laetoli, Olduvai, and Peninj hominin sites. Quaternary International 322: 264 276. Bates, K. T., Rarity, F., Manning, P. L., Hodgetts, D., Bernat, V., Oms, O., Galobart, A., Gawthrope, R. L. 2008. High resolution LiDAR and phot ogrammetric survey of the Fumanya dinosaur tracksites (Catalonia): implications for the conservation and interpretation of geological heritage sites. Journal of the Geological Society, London 165: 115 127. Behrensmeyer, A.K. 2006. Climate Change and Hum an Evolution. Science 311, 476 478 Behrensmeyer, A.K and Reed, K.E., 2013. Reconstructing the Habitats of Australopithecus: Paleoenvironments, Site Taphonomy, and Faunas. In: Reed, K.E., Fleagle, J.G., Leakey, R.E., editors. The Paleobiology of Australopithecus. Vertebrate Paleobiology and Paleoanthropology. Springer Netherlands 41 60. Bennet, M. R., Falkingham, P., Morse, S.A., Bates, K., Crompton, R. H. 2013. Preserving the Impossible: Conservation of Soft Sediment Hominin Footprint Sites and Strategies for Three Dimensional Digital Data Capture. PLoS One 8 (4): e60755. Bennet, M. R. and Morse, S. A. 2014. Human Footprints: Fossilised Locomotion? Springer International Publishing Bennet, M. R., Reynolds, S. C., Morse, S. A., Budka, M. 2016. Laetoli's lost tracks: 3D generated mean shape and missing footprints. Nature Scientific Reports 6: e21916. Bennet, M. R., Reynolds, S. C., Morse, S. A., Budka, M. 2016. Footprints and human evolution: Homeostasis in foot function? Paleogeography, Paleoclimatology, Paleoecology 461: 214 223. Berge, C., Penin, X., Pelle, E. 2006. New interpretation of Laetoli footprints using an experimental approach and Procrustes analysis: Preliminary resu lts. C.R. Palevol 5: 561 569.

PAGE 83

74 Bobe, R., Behrensmeyer, A.K., Chapman, R.E., 2002. Faunal change, environmental variability and late Pliocene hominin evolution. Journal of Human Evolution 42: 475 497. Brunet, M., Beauvilain, A., Coppens, Y., Heintz, E., Moutaye, A.H.E., Pilbeam, D.1996. Australopithecus bahrelghazali de Koro Toro (Tchad). Sciences de la terre et des plantes. 322, 907 913. Brunet, M., Guy, F., Pilbeam, D., Mackaye, H. T., Likius, A., Ahounta, D., Beauvilain, A., et al. 2002. A new hominin from the Upper Miocene of Chad, Central Africa. Nature 418:145 151. Clarke, R., 2013. Australopithecus from Sterkfontein Caves, South Africa. In: Reed, K.E., Fleagle, J.G., Leakey, R.E., editors. The Paleobiology of Australopithecus. Vertebrate Paleobiology and Paleoanthropology. Springer Netherlands 105 123. Crompton, R. H., Pataky, T. C., Savage, R., D'Aout, K. D., Bennet, M. R., Day, M. H ., et al. 2012. Human like external function of the foot, and fully upright gait, confirmed in the 3.66 million year old Laetoli hominin footprints by topographic statistics, experimental footprint formation and computer simulation. Journal of the Royal Society Interface 9: 707 719. DeSilva, J. M. 2010. Revisiting the "Midtarsal Break." American Journal of Physical Anthropology 141: 245 258. DeSilva, J. M., Bonne Aannee, R., Swanson, Z., Gill, C. M., Sobel, M., Uy, J., and Gill, S. V. 2015. Midtars al Break Variation in Modern Humans: Functional Causes, Skeletal Correlates, and Paleontological Implications. American Journal of Physical Anthropology 156: 543 552. Dingwall, H. L., Hatala, K. G., Wunderlich, R. E., Richmond, B.G. Hominin stature, bod y mass, and walking speed estimates based on 1.5 million year old fossil footprints at Ileret, Kenya. Journal of Human Evolution 64: 556 568. Ditchfield, P. and Harrison, T. 2011. Sedimentology Lithostratigraphy and Depositional History of the Laetoli Area. In: Paleontology and Geology of Laetoli: Human Evolution in Context. Harrison, T., Editor. Springer Science+Business Media. Domnguez Rodrigo, M. the Emergenc e of the Earliest Hominins? Current Anthropology 55, 59 81. Foley, R. 2002. Adaptive Radiations and Dispersals in Hominin Evolutionary Ecology. Evolutionary Anthropology Supplement 1: 32 37.

PAGE 84

75 Foley, R. and Gamble, C., 2009. The Ecology of Social Transit ions in Human Evolution. Philosophical Transactions: Biological Sciences 364: 1533, 3267 3279. Grabowski, M. Hatala, K. G., Jungers, W. L., Richmond, B. G. 2015. Body mass estimates of hominin fossils and the evolution of human body size. Journal of Hum an Evolution 85: 75 93. Haile Selassie, Y., Saylor, B.Z., Deino, A., Levin, N.E., Alene, M., Latimer, B.M., 2012. A new hominin foot from Ethiopia shows multiple Pliocene bipedal adaptations Nature 483, 565+. Haile Selassie, Y., Gibert, L., Melilo, S M., Ryan, T. M., Alene, M., Deino, A., Levin, N. E., Scott, G., Saylor, B. Z. 2015. New species from Ethiopia further expands Middle Pliocene hominin diversity. Nature 521: 483 488. Haile Selassie Y. 2016. Introduction to KSD VP 1/1: The Earliest Ad ult Partial Skeleton of Australopithecus afarensis In: Haile Selassie Y., Su D.F., editors. The Postcranial Anatomy of Australopithecus afarensis. Springer Netherlands 1 12. Haile Selassie Y., Melillo S.M., Su D.F. 2016. The Pliocene hominin diversit y conundrum: Do more fossils mean less clarity? PNAS 113: 6364 6371. Hatala, K. G., Demes, B., Richmond, B. G. 2016. Laetoli footprints reveal bipedal gait biomechanics different from those of modern humans and chimpanzees. Proceedings of the Royal S ociety B 283: 20160235. Harcourt Smith, W.E.H. and Aiello, L.C., 2004. Fossils, feet and the evolution of human bipedal locomotion. Journal of Anatomy 204, 403 416. Harrison, T. 2011. Laetoli Revisited: Renewed Paleontological and Geological Investiga tions at Localities on the Eyasi Plateau in Northern Tanzania. In: Paleontology and Geology of Laetoli: Human Evolution in Context. Volume 1: Geology, Geochronology, Paleoecology and Paleoenvironment. Harrison, T., Editor. Springer Science+Business Me dia. 1 15. Harrison, T. 2011. Hominins from the Upper Laetolil and Upper Ndolanya Beds, Laetoli. In: Paleontology and Geology of Laetoli: Human Evolution in Context, Volume 2: Fossil Hominins and Associated Fauna. Harrison, T., Editor. Springer Science +Business Media. 141 188. Hay, R. L. 1987. Geology of the Laetoli Area. In: Laetoli: A Pliocene Site in Northern Tanzania. Leakey, M. D. and Harris, J. M., (Editors). Clarendon Press Oxford Hunt, K. D. 1994. The evolution of human bipedality: ecology and functional morphology. Journal of Human Evolution 25: 183 202.

PAGE 85

76 Johanson, D. and Edey, M.A., 1990. Lucy: The Beginnings of Humankind. Simon and Schuster. Johanson, D. C. 2004. Lucy, Thirt y Years Later: An Expanded View of Australopithecus afarensis. Journal of Anthropological Research 60 (4): 465 486. Johanson, D. C., Lovejoy, C. O., Kimbel, W. H., White, T. D., Ward, S. C., Bush, M. E., Latimer, B. M. Coppens, Y., 1982. Morphology of the Pliocene Partial Hominid Skeleton (AL 288 1) from the Hadar Formation, Ethiopia. American Journal of Physical Anthropology 57, 403 451. Kim, J. Y., Kim, K. S., Lockley, M. G., Matthews, N. 2008. Hominid Ichnotaxonomy: An Exploration of a Neglected Discipline. Ichnos 15: 26 139. Kimbel, W.H., Lockwood, C.A., Ward, C.V., Leakey, M.G., Rak, Y., Johanson, D.C., 2006. Was Australopithecus anamensis ancestral to A. afarensis? A case of anagenesis in the hominin fossil record. Journal of Human Evolut ion. 51, 134 152. Kimbel, W. H., and Delezene, L. K. 2009. "Lucy" Redux: A Review of Research on Australopithecus afarensis Yearbook of Physical Anthropology 52: 2 48. Kingston, J. D. and Harrison, T. 2007. Isotopic dietary reconstructions of Pliocene herbivores at Laetoli: Implications for early hominin paleoecology. Paleogeography, Paleoclimatology, Paleoecology 243: 272 306. Klenerman, L., and Wood, B. 2006. The Human Foot: A Companion to Clinic Studies. Springer Verlag London Limited Kuhn, S. L., Raichlen, D.A., Clark, A.E., 2016. What moves us? How mobility and movement are at the center of human evolution. Evolutionary Anthropology: Issues, News, and Reviews 25, 86 97. Latimer, B. and Lovejoy, C.O., 1989. The calcaneus of Australopithecus afarensis and its Implications for the evolution of bipedality. American Journal of Physical Anthropology 78, 369 386. Latimer, B. and Lovejoy, C.O., 1990. Metatarsophalangeal joints of Australopit hecus afarensis. American Journal of Physical Anthropology 83, 13 23. Leakey, M.D. and Hay R.L. 1979. Pliocene footprints in the Laetolil Beds at Laetoli, northern Tanzania. Nature 278, 317 323. Leakey, M.D., 1981. Discoveries at Laetoli in northern Tanzania. Proceedings of the 92, 81 86.

PAGE 86

77 Leakey, M.D. 1987. The Hominid Footprints: Introduction. In: Laetoli: A Pliocene Site in Northern Tanzania. Leakey, M. D. and Harris, J. M., (Editors). Clarendon Press Oxford Leakey, M. D. and Harris, J. M., (Editors). 1987. Laetoli: A Pliocene Site in Northern Tanzania. Clarendon Press Oxford. Leakey, M. G., Spoor, F., Brown, F. H., Gathogo, P. N., Klarle, C., Leakey, L. N., McDo ugall, I. 2001. New hominin genus from eastern Africa shows diverse middle Pliocene lineages. Nature 410: 433 440. Lockley, M. G. 1998. The Vertebrate Track Record. Nature 396: 429 433. Lovejoy, C. O. 1981. The Origin of Man. Science 211 (4480): 341 350. Lovejoy, C.O., 2009. Reexamining Human Origins in Light of Ardipithecus ramidus Science. 326, 74+ Lovejoy, C.O., Latimer, B. M., Spurlock, L., Haile Selassie, Y. 2016. The Pelvic Girdle and Limb Bones of KSD VP 1/1. In: Haile Sela ssie Y., Su D.F. editors. The Postcranial Anatomy of Australopithecus afarensis. Springer Netherlands 155 178. Masao, F.T., Ichumbaki, E.B., Cherin, M., Barili, A., Boschian, G., Iurino, D.A., Menconero, S., Moggi Cecchi, J., Manzi, G. 2016. New footprints from Laetoli (Tanzania) provide evidence for marked body size variation in early hominins. eLife 5, e19568. Matthews, N. A. 2008. Aerial and Close Range Photogrammetric Technology: Providing Resource Documentation, Interpretation, and Preservation. Technical Note 428. U.S. Department of the Interior, Bureau of Land Management, National Operations Center, Denver, Colorado. Matthews, N., Noble, T., Breithaupt, B. 2016. Close Range Photogrammetry for 3 D Ichonology: The Basics of Photogrammetric Ichnology. In: Falkingham, P. L., Marty, D., Richter, A., eds. Dinosaur Tracks: The Next Steps. Indiana University Press. McHenry, H. M. 1986. The First Bipeds: A Comparison of the A. afarensis and A. africanus Postcranium and Implicati ons for the Evolution of Bipedalism. Journal of Human Evolution 15: 177 191. McHenry, H.M., 1991. Sexual dimorphism in Australopithecus afarensis. Journal of Human Evolution 20, 21 32. McHenry, H. M. 1994. Early Hominid Postcrania Phylogeny and Func tion. In: Integrative Paths to the Past: Paleoanthropological Advances in Honor of F. Clark Howell. Corruccini, R. S. and Clochon, R. L., editors. Prentice Hall 250 268.

PAGE 87

78 McHenry, H. M., Jones, A.L., 2006. Hallucial convergence in early hominids. Journal of Human Evolution 50, 534 539. McHenry, H.M. and Brown, C.C. 2008. Side steps: the erratic pattern of hominin postcranial change through time. Journal of Human Evolution 55: 639 651. Meldrum, D. J., Lockley, M. G., Lucas, S. G., Musiba, C. 2 011. Ichnotaxonomy of the Laetoli trackways: The earliest hominin footprints. Journal of African Earth Sciences 60 1 12. Musiba, C.M., Tuttle, R.H., Hallgrmsson, B., Webb, D.M., 1997. Swift and sure footed on the Savanna: A study of Hadzabe gaits and f eet in Northern Tanzania. American Journal of Human Biology 9, 303 321. Musiba, C. M., Magori, C., Stoller, M., Stein, T., Branting, S., Vogt, M., Tuttle, R., et al. 2007. Taphonomy and paleoecological context of the Upper Laetolil Beds (Localities 8 a nd 9), Laetoli in northern Tanzania. In: Bobe, R., Alemseged, Z., Behrensmeyer, A. K., eds. Hominin Environments in the East African Pliocene: An Assessment of the Faunal Evidence. Springer Netherlands. 257 278. Musiba, C. M., Mabula, A. Selvaggio, M. Magori, C. C. 2008. Pliocene Animal Trackways at Laetoli: Research and Conservation Potential. Ichnos 15: 166 178. Nathan, R., Getz, W. M., Revilla, E., Hoyoak, M., Kadmon, R., Saltz, D., Smouse, P.E. 2008. A movement ecology paradigm for unifying org anismal movement research. PNAS 105 (49): 19052 19059. Pickford, M., 2006. Paleoenvironments, Paleoecology, Adaptations, and the Origins of Bipedalism in Hominidae. 175 198. Plavcan, J.M., Lockwood, C.A., Kimbel, W.H., Lague, M.R., Harmon, E.H., 2005. Sexual dimorphism in Australopithecus afarensis revisited: How strong is the case for a human like pattern of dimorphism? Journal of Human Evolution 48, 313 320. Potts, R., 1994. Variables versus models of early Pleistocene hominid land use. Journal o f Human Evolution 27, 7 24. Potts, R., 1998. Variability Selection in Hominid Evolution. Evolutionary Anthropology 7, 81 96. Potts, R., 2007. Environmental Hypotheses of Pliocene Hominin Evolution. In: Bobe, Alemseged, and Behrensmeyer (eds.) Hominin Environments in the East African Pliocene: An Assessment of the Faunal Evidence 25 49. Potts, R., 2012. Environmental and Behavioral Evidence Pertaining to the Evolution of Early Homo. Current Anthropology 53:56, S299 S317.

PAGE 88

79 Prang, T.C., 2015. Calcaneal robusticity in Plio Pleistocene hominins: Implications for locomotor diversity and phylogeny. Journal of Human Evolution 80, 135 146. Raichlen, D. A., Pontzer, H., Sockol, M. D. 2008. The Laetoli footprints and early hominin locomotor kinematics. Journal of Human Evolution 54: 112 117. Raichlen, D. A., Gordon, A. D., Harcourt Smith, W. E.H., Foster, A. D., Haas Jr., W. R. 2010. Laetoli Footprints Preserve Earliest Direct Evidence of Human Like Bipedal Biomechanics. PLoS One 5 (3): e9769. Reed, K.E., 1997. Early hominid evolution and ecological change through the African Plio Pleistocene. Journal of Human Evolution 32, 289 322. Reno, P. L. McCollum, M. A., Meindl, R. S., Lovejoy, C. O. 2010. An enlarged postcranial s ample confirms Australopithecus afarensis dimorphism was similar to modern humans. Philosophical Transactions of the Royal Society B 365: 3355 3363. Richmond, B.G., Jungers, W.L., 1995. Size variation and sexual dimorphism in Australopithecus afarensis and living hominoids. Journal of Human Evolution 29, 229 245. Robbins, L.M. 1987. Hominid footprints from Site G. In: Laetoli: A Pliocene Site in Northern Tanzania. Leakey, M. D. and Harris, J. M., (Editors). Clarendon Press Oxford Ruff, C., 2002. V ariation in Human Body Size and Shape. Annual Review of Anthropology 31, 211 232. Sayers, K. and Lovejoy, C. O., 2008. The Chimpanzee Has No Clothes: A Critical Examination of Pan troglodytes in Models of Human Evolution. Current Anthropology 49:1, 8 7 114. Stern, J. T., and Susman, R. L., 1983. The Locomotor Anatomy of Australopithecus afarensis. American Journal of Physical Anthropology. 60, 279 317. Su D. F. and Harrison T. 2008. Ecological implications of the relative rarity of fossil hominin s at Laetoli. Journal of Human Evolution 55:672 681. Tattersall, I., 1986. Species recognition in human paleontology. Journal of Human Evolution 15, 165 175. Tuttle, R. H., 1987. Kinesiological inferences and evolutionary implications from Laetoli bip edal trails G 1, G/23, and A. In: Leakey, M. D. and Harris, J. M., Editors. Laetoli: A Pliocene Site in Northern Tanzania. Clarendon Press Oxford. 503 523. Tuttle, R. H., Webb, D., Weidl, E., Baksh, M. 1990. Further Progress on the Laetoli Trails. Jou rnal of Archaeological Science 17: 347 362.

PAGE 89

80 Tuttle, R. H., Webb, D.M., Baksh, M. 1991. Laetoli Toes and Australopithecus afarensis Human Evolution 6 (3): 193 200. Tuttle, R. H., Webb, D. M., Tuttle, N. I. 1991. Laetoli Footprint Trails and the Evolution of Hominid Bipedalism. Origine(s) De La Bipedie Chez Les Hominides. Paris: Cahiers de Paleioanthropologie, Editions du CNRS : 203 221. Tuttle, R. H. 2008. Footp rint Clues in Hominid Evolution and Forensics: Lessons and Limitations. Ichnos 15, 158 165. Ward, C. V., 2002. Interpreting the Posture and Locomotion of Australopithecus afarensis: Where Do We Stand? Yearbook of Physical Anthropology. 45, 185 215. W ard, C. V., Kimbel, W. H., Harmon, E. H., Johanson, D.C., 2012. New postcranial fossils of Australopithecus afarensis from Hadar, Ethiopia (1990 2007). Journal of Human Evolution. 63, 1 51. Ward, C.V., Kimbel, W.H., Johanson, D.C., 2011. Complete Fourt h Metatarsal and Arches in the Foot of Australopithecus afarensis. Science. 331, 750 753. Ward, C.V., 2013. Postural and Locomotor Adaptations of Australopithecus Species. In: Reed, K.E., Fleagle, J.G., Leakey, R.E. (Eds.), The Paleobiology of Austra lopithecus, Vertebrate Paleobiology and Paleoanthropology. Springer Netherlands 235 245. White, T.D. and Suwa, G., 1987. Hominid Footprints at Laetoli: Facts and Interpretations. American Journal of Physical Anthropology. 72, 485 514. Zaitsez, A. N., Wenzel, T., Spratt, J., Williams, T. C., Strekopytov, S., Sharygin, V. V., et al. 2011. Was Sadiman volcano a source for the Laetoli Footprint Tuff? Journal of Human Evolution 61: 121 124. Zipfel, B., DeSilva, J. M., Kidd, R. S., Carlson, K. J., Churc hill, S. E., Berger, L. R. 2011. The Foot and Ankle of Australopithecus sediba. Science 333: 1417 1420.