MORPHOLOGICAL AFFINITIES OF RECENTLY DISCOVERED CERCOPITHECIDS FROM THE PLIOCENE UPPER LAETOLIL BEDS IN LAETOLI, TANZANIA by ELICIA F. ABELLA B.S., The Pennsylvania State University, 2012 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 201 4
ii 2014 ELICIA F. ABELLA ALL RIGHTS RESERVED
iii This thesis for the Master of Arts degree by Elicia Frances Abella has been approved for the Anthropology Program by Charles Musiba, Chair Zaneta Thayer Tammy Stone May 2, 2014
iv A bella, Elicia F. (M.A Anthropology) Morphological Affinities of Recently Discovered Fossil Cercopithecids from the Pliocene Upper Laetolil Beds in Laetoli, Tanzania Thesis directed by Associate Professor Charles Musiba ABSTRACT The Laetoli paleoanthropological site in northern Tanzania continues to yield one of the oldest Australopithecus afarensis collection as well as other well preserved non hominin primate remains including primate cercopithecids. The Laetoli primat e s are highly diversified, including fossil galagines, parapithecids and paracolobines. These primate species are indicative of highly variable depositional environments at Laetoli that would have been more wooded or forested with patches of bushes, thorn scrubs and op en habitats. For example, extant cercopithecines tend to thrive in a wide range of habitats from grasslands to dense woodlands, forests and tree covered rocky outcrops. This thesis presents qualitative descriptions and a computational statistical analysis of odontometrics for the posterior teeth of fossil and extant primates in order to classify recently recovered Laetoli fossil primate remains by the University of Colorado Denver Tanzania field s chool in paleoanthropology The results of this analysis wil l help give a better understanding of the ecological diversity exploited at Laetoli. The form and content of this abstract are approved I recommend its publication. Approved by: Charles Musiba
v TABLE OF CONTENTS CHAPTER I. INTRODUCTION ...... ..1 II. BA CKGROUND RESEARCH ... 3 Theoretical B ackground 3 Current L iterature: Geology and Paleoecology .. ... 4 Current L iterature: Primates ..11 III. MATERIALS AND METHODS .......................................... .................................20 Sample 20 Craniofacial D escriptions ... 22 Methods: Dental M etrics 37 Methods: Statistical Analysis 42 IV. RESULTS .. 4 4 P 3 44 P 4 ... 46 M 1 .. 48 M 2 50 M 3 .. 52 Summary: Results ..... 54 V. DISCUSSION ..55 VI. CONCLUS 66 REFERENCES .68
vi APPENDIX A. Summary of Collected Cercopithecine and Colobine Odontometrics B. Summary of Eigenvalues from the PCA for P 3 M 3 ............................... ............. ...72 C. Summary of Percentage of Total Variance from the PCA for P 3 M 3 ................. ............................ .... ................. ........................... ... 72 D. Summary of the DFA R esults for P 3 M 3 .................... ................... ................... ......72 E. Summary of the C orrect ly Classified and M isclassified P ercentages from the DFA for P 3 M 3 .................... ............................... .................72
vii LIST OF TABLES TABLE 1. Summary of Cercopithecines and Colobines used in the quantitative and qualitative 2. Method of mandibular odontometr ics taken for proposed Cerco pithecid specimens from Laetoli (Swindler 2002) .. .. 38 3. Mandibular odontometrics from the proposed Cercopithecid sp. from Laetoli, Local 2 and LP 061703 01&02 fragment; right mandibular segment from P 3 M 3 in .. .41
viii LIST OF FIGURES FIGURES 1. Geological map of Laetol 2. Map of the Laetoli area and the 3. Figure 3: Sections showing the sequence exposed in Loc. 6 ............................ .....10 4. Cladogram of extant cercopithecid genera and subgenera............................................................... ..............................19 5. Left maxillary M 1 M 3 fragment of the Laetoli cercopith from Locality 6...................................... ..............................23 6. Right maxillary P 3 P 4 fragment of the Laetoli cercopith from Locality 8...................................... ..............................25 7. Left ma xillary M 1 of the Laetoli cf. Rhinocolobus sp. (specimen 04 1 01) from Locality 1 ........................................................................ ..............................28 8. Right mandibular M 1 M 2 fragment of Parapapio ado from Locality 1........................................... ..............................31 9. Mandibular specimen of Parapapio ado from Locality 2; occlusal view................................................ ..............................35 10. Mandibular specimen of Parapa pio ado from Locality 2; lateral view.................................................... ..............................35 11. Comparative mandibular specimens for the Parapapio ado Laetoli specimen from Locality 2........................................................................ ..............................36 12. Odontometric landmarks................................................................ ........................39
ix 13. Identification of cerco pithecid dental landmarks on the right mandibular M 3 of Theropithecus .................................................................................................. .40 14. 3D score scatterplot for the P 3 PCA.................................................. ..... ................45 15. 3D score scatterplot for the P 4 PCA.................................................. ...... ............... 47 16. 3D score scatterplot for the M 1 PCA................................. ............... ...... ............... 49 17. 3D score scatterplot for the M 2 PCA................................................... ...... ............ 51 18. 3D score scatterplot for the M 3 PCA...................................................... ... ............ 53
1 CHAPTER I INTRODUCTION Laetoli, one of the most important paleontological and paleoanthropological sites located in northern Tanzania, has yielded the second largest sample of Australopithecus afarensis specimens and is also famous for its preserved trails of hominin footprints (Su & Harrison 2007). However, having a better understanding of the paleobiology of Australopithecus afarensis continues to be one of the mo st controversial topics to day. For example, previous interpretations of the paleoecology at Laetoli suggest that Aus tralopithecus afarensis inhabited an arid to semi arid grassland with scattered bush and tree cover containing patches of woodland which are all characteristics of present day Serenge ti Plains (Leakey et al. 1976). However, studies by Andrews (1987), Reed (1997) and Musiba (1999) have also suggested that Laetoli had more of a dense bush cover and more extensive tracts of woodland and forest galleries. Reed (1997) suggests that based on the high taxonomic diversity of mammalian fossil assemblages, which incl ude arboreal and frugivorous mammals, Laetoli must have had significant components of bushland and woodland. The existence of more than three species of cercopithecines and paracolobines also strongly suggest the existence of forested environments during t he Pliocene at Laetoli (Andrews 1987). Mammalian fauna f ound at Laetoli further support the notion that Laetoli must have had more wooded, bush, and thorn scrub habitats than present day. Laetoli primate species including, Galago sadimanensis, Parapapio a do, and Paracolobus sp., is
2 suggestive of a closed woodland environment or forest; though, cercopithecids do tend to presently thrive in a range of habitats from grasslands to forests, stands of trees and rocky outcrops as common sleeping quadrants. Even though rocky outcrops are not generally seen in Laetoli, larger trees would have been important sites of ref u ge for cercopithecid primates (Su & Harrison 2007) In order to reconstruct an accurate paleoenvironment of Laetoli, a better understanding of the Laetoli primates m ust be considered. By utilizing odontometrics in a statistical analysis with qualitative de scriptions and comparisons between fossil and extant African primates, determining the species of the recently recovered Laetoli primates is suggestive of their depositional envir onments and dietary adaptations during the Pliocene. Qualitative de scriptions and comparisons help support the quantitative measures to distinguish primate species. Specific aims of this thesis include: 1. Establishing a comparative odontometric data set on extant and fossil Cercopithecids (subfamily: Cercopithecinae and Colobinae) fou nd in Africa. 2. Investigating primate species from Laetoli to better understand their habitats during the Pliocene. 3. Utilizing Principle Component Analysis (PCA) and Discriminant Function Analysis (DFA) of odontometrics to differentiate primate species from L aetoli.
3 CHAPTER II BACKGROUND RESEARCH Theoretical B ackground Evolutionary ecology suggests that behaviors are adaptive in terms of environmental variability in which an organism is able to thrive by enhancing their reproductive success. However, Behavioral Ecology (BE) focuses on behaviors as a phenotypic result of natural selection favoring adaptive options that have the ability to solve fitness related trade offs (Bird & ontogenetic and other exter nal factors that may be relatable to the cost and benefits of certain behaviors. By using the environment of evolutionary adaptedness (EEA), it can refer to the selective environment in which an individual thrives. Here, the EEA creates a functional link b etween the cognitive systems and the context in which they have evolved (Foley 2005). In order to apply the EEA to paleoanthropology, Foley (2005) suggests that speciation is best analyzed by considering distinctive morphologic characteristics within the appropriate comparative biological framework. Utilizing odontometrics and its functional biomechanical capabilities, this research distinguishes species within Colobinae and Cercopithecinae from the Upper Laetolil Beds. In turn, understanding the sequence and accumulation of distinctive odontometric traits can establish proximate and ontogenetic factors that influence dietary adaptations in the fossil record (Foley 2005). Incorporating this theoretical framework allows this research to focus on the selected ecological niche a species has and how each species reacts differently to
4 environmental pressures, such as the cercopithecid evolutionary trend for a n expanded diet due to terrestriality. Current L iterature: Geology and Paleoecology Laetoli, located in the southern part of the Eastern Rift Valley, is surrounded by a series of tilted fault blocks with lake basins and plateaus. However, Laetoli can be found on an upthrown fault block to the north of Lake Eyasi and Olduvai Gorge (Hay 1987). The Laetolil Beds were first studied by Kent (1941) who used the Vogel River Series to divide the lower unit (Laetolil Beds) and the upper unit (Ngaloba Beds). The Laetol il Beds were first described as sub aerially deposited tuffs having an irregular surface overlain by agglomerate and nephelinite lava where it separates from the Ngaloba Beds by a grey limestone with bright red pebble like bodies that continues to the southwest of Laetoli (Kent 1941).
5 Figure 1: Geological map of Laetoli (
6 From 1974 to 1979, Hay conducted a stratigraphical investigation that revealed more stratigraphical units in the Laetoli area, such that the Ngaloba Beds are followed b y the Olpiro Beds, Naibadad Beds, Ndolanya Beds and the Laetolil Beds. The Laetolil Beds are then divided into the lower and upper units, where these two units can be distinguishable. The lower unit of th e Laetolil Beds consists mostly of aeolian tuff inte rbedded with air fall and water worked tuff that also contains a few beds of conglomerate and breccia sediments Like the lower unit, the upper unit consists mostly of aeolian tuff but contains some air fall tuffs and a few horizons of xenoliths (Leakey et al. 1976). The Yellow Marker Tuff is a distinctive pale yellow vitiric tuff that forms the topmost stratum of the upper unit and is used to distinguish the Upper Ndolanya Beds from the Upper Laetolil Beds (Hay 1987). The Upper Laetolil Beds consist of the Australopithecus afarensis specimens that are radiometrically dated to Ma (Deino 2011). Furthermore, the Upper Beds are able to be sub divided into marker tuffs from a series of narrow temporal zones (Hay 1987). According the Hay (1987), the faunal assemblage in the Upper Laetolil Beds represents a grassland savanna with areas of thicker vegetation cover where most of the bovid remains indicate an open country fauna. On the other hand, Musiba (1999) suggests that the Laetoli Pliocene environm ent contain s more woodland areas than previously thought (Leakey et al. 1987). Localities 8 and 9 suggest that both grasslands and galleries of woodland were present due to bovid functional morphologies relating to mosaic environments. Although this resear ch focuses on specific localities of the Upper Laetoli l Beds, other reanalysis of the Upper Beds
7 support the notion of a prolonged, consistent record of a mosaic environment during this time period. Although some aspects of the faunal evidence indicates a range of habitats, which includes a more dense bush cover and more extensive tracts of woodland than seen in the region today, a high frequency of grazers and terrestrial mammals coupled with a low occurrence of arboreal and frugivorous mammals indicate t hat the paleoecology of Laetoli is a mosaic habitat that is dominated by grassland and shrubland with some areas of woodland (open medium and closed woodlands) and gallery forests along seasonal rivers (Su & Harrison 2007). Galigo sadimanensis and at least three species of cercopithecids ( Parapaio ado, Paracolobus sp., and a n unspecified larger colobine monkey) are suggestive of a closed woodland, f orest or rocky outcrop ; however reconstructions of the landscape at Laetoli show an absence of ro cky outcrops implying that trees were the main site of refuge for these primates (Su & Harrison 2007). Although extant species of these primates do occupy a range of habitats such as grasslands and forests, other mammals ( Paraxerus, Rhychocyon, Subulona, E uonyma) can be used in the reconstruction of Laetoli to show its mosaic environment (Su & Harrison 2007). The current paleoecolog ical reconstructions of the Upper Laetolil Beds show that this time period represents a heterogeneous mosaic environment of gra ssland, savanna h and woodland habitats. Su & Harrison (2007 ) also found that there were no significant differences in ecological diversity between the different localities or stratigraphic zones in the Upper Laetoli l Beds. This implies that the composition of the mammalian fauna was essentially identical throughout the entire Upper Laetolil sequence and a consistent ecological structure was maintained throughout
8 this time regardless of regional or local environmental di sturbances (Su & Harrison 2007 ). Howe ver, their study suggests that the mosaic environment of the Upper Laetolil sequence was relatively stable, rather than a mixture of time averaged habitats. The specimens being examined in this research are derived from Locality 1 (between T uff 6 and Yellow Marker Tuff), Locality 2, Locality 6 (between Tuffs 5 and 7) and Locality 8 Figure 2: Map of the Laetoli area and the location of localities (taken from Dr. Charles Tuff 6 from Locality 1 c onsists mostly of a brown tefra with a clayey sand consistency that contains mostly aeolian tuff with patches of xenoliths; whereas T uff 7 and 8 consists mostly of grey clayey sand tefra that contains a patch of xeno liths near the
9 topmost part of T uff 8 Tuff 9 and the Yellow Marker T uff from Lo cality 1 contains mostly yellow grey colored sediment where it is compri sed mostly of air fal l tuff and a water wo rked tuff toward the bottom of Tuff 9. However, present in T uff 9 is an orange colored air fall tuff that can b e seen on the upper segment of T uff 9 (Hay 1987; see Fig. 2.1 and Fig. 2.3) According to Hay (1987), the most fossiliferous tuffs at Locality 1 are found between Tuff 6 and the Yellow Marker Tuff, which is consistent with the findings of the Laet oli primate specimen s being examined Locality 1 is situated in the northeast section of Laetoli, which is also northeast of Locality 23 (Hay 1987; see Fig. 2.9). Tuff 5 to T uff 7 from Locality 6 contains a dominantly grey, clayey sand that is mostly of ae olian tuff with patches of water worked tuffs and xenoliths. However, toward the lower end of T uff 5 is a yellow and yellow grey aeolian tuff which is overlain by a brown clayey, aeolian tuff (Hay 1987; see Fig. 2.1 and Fig. 2.3). Hay (1987) also suggests that the most fossiliferous tuffs in Locality 6 are found between Tuff 5 and Tuff 6, which is also consistent with the findings presented Locality 6 is situated on the mid eastern section of Laetoli, north of Locality 7 and east of the Laetoli camp site (Hay 1987; see Fig. 2.9).
10 Figure 3: Sections showing the sequence exposed in Loc. 6 Note the stratigraphic gaps in the section are not shown to sca le. (Taken from Ditchfield and Harrison 2011; See Fig 3.10 pg. 58)
11 According to Su and Harrison (2008), the fossil cercopithecid assemblage found at Laetoli consist s mostly of cranio dental specimens; a pattern also similar to the rarity of fossil hominins. They also suggest that this pattern is not only seen in hominins and cercopithecids, but in large mammals found at Laetoli. Their findings suggest that the craniodental remains are significantly overrepresented compared with postcranial elements and that a phenomenon related to broader taphonomic factors must be occurring at Laetoli. Although teeth are the most resilient skeletal structures in the mammalian skeleton, they sugge st tha t the disproportionality is due to carnivore scavenging and periodical ash falls at Laetoli (Su & Harrison 2008). Current L iterature : Primates In 1976, Delson conducted a systematic study pertaining to the classification of subgroups of cercopithecids with the use of cranial morphologies and factor analysis. His findings showed that the primary distinction within Cercopithecidae can be determined between the lo nger face cercopithecines and the shorter more upright faced colobines. Given its overall cranial morphologies, the facial height is much greater in cercopithecines in the suborbital zygomatic region when compared to colobines, whereas colobines have a rel atively wider facial region (Delson 1976). Because of these facial and cranial morphological differences between cercopithecines and colobines, the dental morphology will relate to such anatomical changes in the craniofacial region, including the mandible. Typically, ce rcopithecid dentition involves molariform teeth with asymmetrically high crowns with four m arginal cusps connected by lophid s (transverse ridges) and three foveas separated by two ridges. The maxillary teeth are usually mirror images of the
12 mandibular teeth where the buccal and lingual characteristics are reversed (Szalay and Delson 1979). On the other hand, the mandible in colobines and cercopithecines possess specialized characteristics for each subfamily, indicative of specializations in masticatory functions. Colobine mandibles tend to have a rel atively upright ascending ramus; whereas most cercopithecines have an extremely back tilted ramus, especially found in larger papionins. However, all colobines possess an expan sion of the gonial angle with inferior bulging beneath the rear molars which corresponds to its overall deeper corpus. The corpus in colobines also tend to possess a relatively constant height and a mesially shallowing corpus, whereas cercopithecines tend to have an increasi ngly mesial depth which is shared by other hominoids (Delson 1976). Cercopithecids have an increased relief of buccal teeth, producing longer marginal shearing crests, which is indicative of folivory. However since cercopithecids are terrestrial animals, the tendency to maintain folivory is unlikely and a broadening diet toward omnivorous foods is attributed to the expansion into the range of habitats that cercopithecids occupied by the Plio Pleistocene (Szalay and Delson 1979). Features such as a reduced cingulum with increased crown relief, an increase of mandibular molar trigonid length with a reduction of the hypoconulid, and an elongation of the maxillary premolars and molars, all exhibit signs of morphological evolution toward omnivory for cercopithe cids (Szalay and Delson 1979). Cercopithecids were first discovered at Laetoli in 1935 by L.S.B. Leakey; however more fossil cercopithecids have been discovered since then. A year later,
13 Hopwood (1936) described Cercocebus ado from Kenya, which is the ho lotype for this Pliocene species. Following these integral cercopithecid finds, Kohl Larsen made an extensive vertebrate collection in the Laetoli region, including 38 cercopithecids, where most of his finds were found in the Upper Laetolil Beds (Harrison 2011). Futhermore, Dietrich (Harrison 2011:Dietrich 1942) erected a new cercopithecine species within Kohl Papio (Simopithecus) serengetensis assuming that this species was distinct from the holotype Cercocebus ado specimen. Finally, from 1959 to 1964, L.S.B. Leaky and M.D. Leakey recovered additional cercopithecid material from Laetoli, where most of the specimens were collected from Localilty 10 (Localities 10, 10W and 10E) (Harrison 2001). Mary Leakey discovered 81 cranio dental sp ecimens from Laetoli expeditions spanning from 1974 to 1979, where she and Delson (1987) recognized four cercopithecids Parapapio ado, Papionini gen. et sp. indet., cf. Rhinocolobus sp., and Cercopithecoides sp. mostly recovered from a wide variety of localities from the Laetoli and Ndolanya Beds (Leakey & D elson 1987). Harrison then recovered an additional 212 cranio dental specimens and 25 postcranial specimens of cercopithecids at Laetoli spanning from 1998 to 2005, which were all found in the Upper Laetolil Beds except for a Parapapio incisor from the Upper Ndolanya Bed at Locality 7E. Galago sadimanensis is represented by a number of partial mandibles found in the Upper Laetoli l Beds (Walker 1987). Howeve r, the fossil record of galagids from the Plio Pleistocene is relatively poor and the only other extinct species found is Otolemur howelli which is found in the Omo Valley, Ethiopia (Wesselman 1984). Another galagid species was discovered in 2003 from Loc ality 10W in the Upper Laetoli l Beds and is the
14 most complete specimen of a Pliocene galagid (Harrison 2011). Although Galago sadimenensis is the only galagid found from Laetoli, this specimen indicates that it may have been a primitive sister taxon to cro wn galagids (Harrison 2010). Due to the scarcity of fossil galagids, not much can be deduced about the paleoecology of Laetoli; however, Galago senegalensis can be found in riverine and open acacia woodland suggesting that the vegetation at Laetoli during the Pliocene included at least open woodland or thorn scrub (Harrison 2010). Harrison recently collected an additional 83 dental specimens of cercopithecids from the Upper Laetolil Beds, except for a deciduous I 1 of Parapapio ado from the Upper Ndolanya Beds where most of his findings were recovered from numerous localities throughout the Laetoli area (Harrison 2011; see Table 6.1). Recovered were Parapaio ado, cf. Rhinocolobus sp., Cercopithecoides sp., and Papionini gen et. ep. indet. craniodental iso lated teeth specimens. Although most of his findings constituted isolated fragments, no partial or complete crania and mandibles were recovered (Harrison 2011). Most of the current literature suggests that Galagids and Cercopithecids thrived in the Laeto li environment during the Pliocene era; however the number of species within Galagids and Cercopithecids seem to be inconclusive and many are lumped into certain species despite their morphological variation (Walker 1987; Jablonski 1994; Frost 2001; Frost & Delson 2002; Harrison 2011) This research investigates odontometric data and its biomechanical implications of craniodental cercopithecid specimens from the Upper Laetolil Beds in order to provide more insight on its paleoecology. This research focuse s on a systematic method in which speciation can be distinguished bas ed on quantitative odontometric data of P 3 P 4 M 1 M 2 and M 3 tooth types and the qualitative characteristics
15 of its biomechanical capabilities. This suggests that odontometric data will support and focus on the biomechanical morphologies that indicate certain dietary adaptations. In addition to, this research describes unclassified primate specimens found at Localities 1, 6, and 8 from Laetoli. Extinct large colobines can be found across eastern Africa (Ethiopia, Kenya, and Tanzania) during the Plio Pleistocene epoch; however their distribution throughout eastern Africa is contained in a diversified environment (Leakey 1982). A fossil assemblage of Rhinocolobus and Paracolobu s taxa have been found in the Omo Valley of Ethiopia consisting of a paleoenvironment that includes a riverine forest that bordered a river, whereas Cercopithecoides is expected to be found at Omo where the paleoenvironment in dicates an open savanna, away from a river (Leakey 1982). Although Koobi Fora has a paleoenvironment similar to Omo, it contains more fluvial deposits indicating perennial rivers with woodland and open country savanna that borders a fluctuating lake (Leake y 1982: Harris 1978). The mosaic environment at Koobi Fora may have been an ideal environment for Plio Pleistocene primates, such that Cercopithecoides williamsi, Cercopithecoides kimeui, Rhinocolobus and Paracolobus indicates both terrestrial and fores ted paleoenvironment s (Leakey 1982). However, since cercopithecines are mostly found at Laetoli and Olduvai Gorg e, the paleoenvironment suggest that this area is predominately a dry open country savanna that contains areas of dense woodland forests (Leakey 1982). The extant cercopithecid species used in this research are derived from various ecologies throughout Africa, ranging from Central, South, West and East Africa.
16 Within cecopithecinae is the gen us Cercocebus, which include C. galeritus, C. agilis, C. chrysogaster and C. torquatus These species are generally found in equatorial western and eastern African regions ranging from Cameroon to Kenya (Wolfheim 1983) The Tana River mangabey ( C. galerit us) thrives primarily in an open canopy gallery forest along the Tana River. Similarly, t he A gile mangabey ( C. agil i s) are forest old world monkeys and thrive in environments such as swamp forests or riparian forests; however, western populations, like the G olden bellied mangabey ( C. galeritus chrysogaster), thrive along riverine forests. Like the others, t he C ollared mangabey ( C. torquatus) is also a moist forest species that typically thrives in mangroves, coastal, ga llery and inland swamp forests (Wolfhe im 1983). Also within cercopithecinae are the papionins, which include the genera Mandrillus, Papio and Theropithcus. There are two species within Manrillus ; D rills (Mandrillus leucophaeus) which thrive in tropical forested environments with open coun trie s and rock outcrops, while M andrills ( Mandrillus sphinx) tend to thrive in tropical forests that border a savanna or in a low elevated river basin with flat, mountainous and plateau terrains (Wolfheim 1983) The genus Papio consists of 5 species: P. anubis P. cynocephalus, P. hamadryas, P. papio and P. ursinus Papio anubis (O live baboon s ) are found throughout Africa and is the most widespread and abundant papionin species. This species thrive in a variety of habitats ranging from semidesert steppe, arid t horn scurb, open grassland with patches of dense scrub, rocky hills and forest savannas; however, they can be found i n areas that are higher, cooler and more humid. In addition to their mosaic preferential habitat, they also thrive in woodlands, gallery fo rests and tropical rainforests. Papio cynocephalus (Y ellow baboons) are usually
17 found in southern and eastern regions of Africa including coastal regions in the east This species primarily live in savanna and woodland habitats, typically perennial grass covered savannas with variable tree densities or a range of forested environments. Papio hamadryas (Hamadryas or S acred heart baboons) are located in the northeast re gion of Africa and southwestern region of the Arabian penins ula where they prefer habitats that contain arid subdesert steppe, dry short grass plains, alpine meadows and montane grasslands (Wolfheim 1983) Papio papio (G uinea baboons) are specifically foun d in western Africa, where they occupy wooded steppes, woodlands, savannas, dry forests and gallery forests. Papio ursinus (Chacma baboons) li ve across southern Africa from the Atlantic coast to the Indian Ocean. This species typically occupy wooded habita ts and open habitats, including subdesert steppe and savannas. Chacma baboons have a wide habitat tolerance where they can occupy temperatures and humidity levels ranging from near desert conditions to mountainous snowy areas. Papionins also tend to use ta ll trees or rocky outcrops as sites for refuge (Wolfheim 1983) The subfamily Colobinae includes 59 species into 10 genera; however, Colobus and Procolobus genera are used in this research. Colobus angeolensis (Angolan Black and White colobus m onkey) are found in equatorial Africa from the Congo River to the coast of the Indian Ocean in present day Tanzania (Wolfheim 1983) This species is forest dwelling and is predominately found in lowland and montane rain forests; however it can be found in forest pat ches in savannas, lakeside forest borders, riparian forests and coastal forests. Within C. angolensis are a few subspecies such as the Tanzani an Black and White colobus m onkey ( C. angolensis palliates) olobus monkey ( C. angolensis sandbergi) which the Tanzanian species tend to live in montane
18 forests and grasslands, alpine bamboo forests and solid stemmed bamboo brush in the highlands of Tanzania. Colobus polykomos (Western Black and White colobus m onkey) occupy along the southern coast of West Africa from Senega l to Nigeria. T hey are a forest species that prefer primarily wet evergreen, moist evergreen, moist deciduous and dry semideciduous forests. Generally, this species can be found in a variety of closed forests formations except tickets and can also be found in patches of riverine forests in savanna environments (Wolfheim 1983) Colobus versus (Olive colobus m onkey s ) occupy western regions of Africa and thrive in many kinds of forest vegetation such as moist semideciduous forest and evergreen rain forests. Unlike the Western Black and Wh ite colobi nes the O live colobines thrive in dense thickets or dense undergrowth s near rivers or swamps. However, Olive colobnes tend to utilize regenerating forests, dry semideciduous forests and riparian forests in savanna environments. Procolobus badius (Red colobus m onkeys) are closely related to Black and W hite colobine s; however they occupy a discontinuous distribution in western, central, and eastern Africa. Like all colobines, they are a for est species that typically live in a moist lowland or gallery forest East African spe cies of Red c olobus monkeys primarily occupy moist, evergreen forests, acacia woodland and woodland evergreen mix tures (Wolfheim 1983)
19 Figure 4 : Cladogram of extant cercopithecid genera and subgenera. Nodes, in numerical order, identify reconstructed morphotypes for: (1) family Cercopithecidae; (2) subfamilies Colobinae and Cercopithecinae; (3) tribes Cercopithecini and Papionini; and (4) subtribes Presbytina and Colobina. Homonoidea is nearest extant sister taxon to Cercopithecidae. Sub generi names enclosed in enclosed in quotation marks to indicate controversy as to its rank. (Strasser & Delson 1987; Figure 1 pp. 82)
20 CHAPTER III MATERIALS AND METHODS Sample The five Laetoli specimens being described and examined were collected duri ng the University of Colorado Denver Tanzania Field School from 1999 2003. These specimens were found at different localities thr oughout the Laetoli area and were found in the Upper Laetolil Beds, dating from Ma (Deino 2011). All of the specimens are craniodental fragments with the exception of a preserved unknown cercopithecid mandible. Due to the fragmented nature of the craniodental specimens, only two of the specimens will be used in this statistical analysis; however a q ualitative description of the all the specimens is presented. A total of 171 cercopithecines and colobines are examined, where 166 are of extant cercopithecids, 3 extinct fossil primates and 2 unknown Laetoli cercopithecids (See Table 1.0). Collecting quan titative data such as mesiodista l length, buccolingual breadth and other mandibular metrics allows an accurate investigation of the type of species found in La etoli during the Pliocene era. Extant primate d ata is obtained from numerous datasets from the Pr imate Research Institute from Kyoto University, Royal Mu seum for Central Africa (RMCA) and the University of Colorado Denver Anthropology Collection ; whereas the fossil cercopithecids data ( Theropithecus and cf. Rhicocolobus) is obtained from Leakey and cercopithecidae Collection (1987), Plio Pleistocene cercopithecids from Kana m East
21 (Harrison & Harris 1995 ), and New Fossil Cercopithecid Remains From the Humpata Plateau, Southern Angola (Jablonski 1994). Table 1 : Summary of the cercopit hecines and colobines used in the quant itative analysis N Cercopithecidae Cercopithecinae Cercopithecini Cercopithecus sp. 2 Papionini Cercocebus 4 galeritus + 1 galeritus agilis + 20 chrysogaster + 10 Mandrillus leucophaeus 2 sphinx 2 Papio anubis* = 5 anubis anubis + 3 anubis (doguera tessellatus) + 22 nigeriae + 1 cynocephalus + 3 C ynocephalus jubilaeus + 1 cynocephalus 10 kindae + cynocephalus lestes + 2 h amadr yas 14 ursinus + 4 Theropithecus baringensis J,L 2 Colobinae Colobus angolensis + 17 angolensis palliatu s + 1 angolenisis sandbergi + 1 polykomos 10 verus 2 Procolobus badius 7 + Odontometric data taken from the Royal Museum for Central Africa Odontometric data taken fromKyoto University Primate Research Institute = Odontometric data taken from the University of Colorado Denver Anthropology Collection J Fossil Theropithecine from Southern Angola (Jablonski 1994) L Fossil Theropithecine from Laetoli (Leakey 1969)
22 Craniofacial D escriptions Order Primates Linnaeus, 1758 Infr aorder Catarrhini Geoffroy, 1812 Superfamily Cercopithecoidea Gray, 1821 Family Cercopithecoidae Gray, 1821 A left maxillary fragment was found during a surface collection of the Upper Laetolil B ed from Locality 6. This fragment contains a well preserved maxillary M 1 M 2 and M 3 of an unknown cercopithecid from Laetoli. P 4 is partially present on the most mesial aspect of the specimen; however, a coronal cross section of both the lingual and buccal roots of P 4 is easily recognizable where the root canal of the lingual root can is preserved. The distal neck of P 4 is present, but there are no recognizable features of cusp morphology. The protocone and hypoconee, which are the mesiolingual and distolingual cusps, are heavily worn on M 1 M 2 and M 3 indicating that the enamel on the lingual aspect is diminished due to at trition. According to Harrison and Harris (1994), Cercocebus is usually distinguishable by having worn down lingual cusps; howeve r, a more extensive analysis must be considered in order to distinguish the specimen. The enamel on the paracone of M 1 and M 2 is slightly worn down, whereas the enamel on the metacone is still intact for both. However, the paracone height on M 3 is signific antly longer than in the metacone with no present wear due to attrition, whereas the enamel on the metacone of M 3 is slightly lost due to attrition. The inferior maxillary alveolar process is present on both the internal and external aspect of the specimen where the body of the maxilla is absent. The internal inferior maxillary border has a curvilinear rounding indicating it is the part of the distal lingual arch of the molar tooth row. The external inferior maxillary border also possesses a curvilinear ro unding with a lateral flaring
23 superior to M 2 which indicates a partial segment of the alveolar process diverging into the inferior aspect of the zygomatic process of the maxilla. Figure 5 : Left maxillary M 1 M 3 fragment of the Laetol i cercopith from Locality 6; A) buccal view B) lingual view C) o ccusal view Order Primates Linnaeus, 1758 Infr aorder Catarrhini Geoffroy, 1812 Superfamily Cercopithecoidea Gray, 1821 Family Cercopithecoidae Gray, 1821 A highly fragmented right mandibular P 3 and P 4 specimen from L ocality 8 was found 50 centimeters NW of LH 5 located just below the T7 layer. The maximum mesiodistal length is of the specimen, including the superior aspect of the alveolar ridge
24 of the mandible, is about 25.47 mm and the maximum buccoli ngual width is 11.35 mm. The most mesial aspect of the specimen contains the distal portion of the alveoli for the canine and the cusps for P 3 and P 4 are complete obliterated due to taphonomic processes. The cervical margin of P 3 and P 4 are contained withi n the superior aspect of the alveolar margin on the buccal and lingual side; however majority of the mandibular body is absent. The superior aspect of t he lingual alveolar margin has somewhat of a vertical slant, indicating the curvilineal lingual aspect o f the mandible. On the other hand, the superior aspect of the buccal alveolar margin is more vertical and does not possess any curvilinear pattern. The inferior view of the specimen contains the apical roots of P 3 and P 4 which are both contained in a calc ified matrix. Due to the high degree of fragmentation, this specimen will not be included in the analysis. However, this specimen lacks diagnostic features and therefore cannot be ascribed to a primate species.
25 Figure 6 : R ight maxillary P 3 P 4 fragment of the Laetoli cercopith from Locality 8; A) occlusal view B) buccal view C) lingual view
26 Order Primates Linnaeus, 1758 Infr aorder Catarrhini Geoffroy, 1812 Superfamily Cercopithecoidea Gray, 1821 Family Cercopithecoidae Gray, 1821 Subfamily Colobinae Jerdon 1867 Genera Colobus Illiger, 1811 Species cf. Rhinocolobus sp. Leakey, 1982 Laetoli specimen LP 04 1 01 was found in the Upper Laetolil Beds from Locality 1. This specimen is an isolated left upper M 1 in which the protocone, hypocone and paracone are visibly present. Although the apical aspect of theses cusps are slightly absent, there are numerous features on this isolated specimen. The protocone and hypocone are similar in height, but possess a difference between the pyramidal shaped protocone a nd conical shaped hypocone. The paracone is also pyramidal in nature, but it shows no sign of attrition; whereas the protocone possesses features attributed to attritional wear. Remnants of the most mesial aspect of the metacone cervical margin are present on the distobuccal aspect. The lingual maxillary root of M 1 is highly preserved; however the distobuccal maxillary root is fragmented and the apical tip is absent. On the other hand, the mesiobuccal root is absent while still retaining a slight portion of the neck l eading into the paracone of M 1 This M 1 isolated tooth is very similar to cf. Rhinocolobus sp. found at Laetoli, sharing characteristics such as its overall square shape in outline with a slight elongation mesiodistally, subequal in size of the protocone and hypocone, higher buccal cusps, a larger paracone than metacone, and a presence of a D shaped mesial fovea that is weakly bordered by a marginal crest (Harrison 2011). These similarities indicate that this M 1 specimen can be qualitatively clas sified as cf. Rhinocolobus from Laetoli. Rhinocolobus turkanensis is principally known to be found in Koobi Fora and Omo Shungura
27 Formations of the Turkana Basin in Kenya and Ethiopia ranging from 3.4 to 1.5 Ma (Harrison 2011). Rhinocolobus tukanensis cont ributes to the diversity of large colobines from the Plio Pleistocene in East Africa, where cf. Rhinocolobus sp. is the second most common cercopithecid found at Laetoli (Harrison 2011).
28 Figure 7 : Left maxillary M 1 of the Laetoli cf. Rhinocolobus sp. (specimen 04 1 01) from Locality 1 ; A) occlusal view B) buccal view C) lingual view D) mesial view
29 Order Primates Linnaeus, 1758 Infr aorder Catarrhini Geoffroy, 1812 Superfamily Cercopithecoidea Gray, 1821 Family Cercopithecoidae Gray, 1821 Subfamily Cercopithecinae Gray, 1821 Tr ibe Papionini Burnett, 1828 Genera Parapapio Hopwood, 1936 Species Parapapio ado Hopwood, 1936 Specimen 061703 It was found on 17 June 2003 from a surface collection. This specimen contains a partial P 4 and an intact M 1 (061703 01) and M 2 (061703 02) that is heavily worn from the buccal aspect. The mesiobuccal root of P 4 is well preserved and borders the most mesi al aspect of the specimen; however the cusp is destroyed and cannot be measured. The lingual aspect of M 1 is partially missing; however the roots are still intact and can be seen from the apical aspect of the specimen. The superior external alveolar proces s is present, but the mandibular corpus is completely absent. On the other hand, the superior internal alveolar process is present towards the mesial dimension and is partially missing in the distal dimension. Since the mandibular corpus is absent, the api cal roots can be seen of P 4 M 1 and M 2 The mesial apical root of P 4 is intact, but the distal root is partially missing. The apical roots of M 1 and M 2 are still intact; however the mesial and distal apical tips of M 2 are slightly absent The protoconid an d hypoconid of M 1 is heavily worn where the enamel on the buccal aspect of the M 1 is diminished due to attrition. The distal portion of M 2 talonid cusp appears to be partially missing; however, the distal root is still intact and is the most distal portion of the specimen itself. M 2 is partially missing the hypoconid and hypoconulid; however the protoconid, metaconid and entoconid can be seen. Although
30 the distal region of the entoconid is partially missing, the enamel on the lingual cusps is not heavily wo rn. The enamel on the buccal cusps, however, are diminished due to attrition.
31 Figure 8 : R ight mandibular M 1 M 2 fragment of Parapapio ado from Locality 1; A) occlusal view B) buccal view C) lingual view
32 Order Primates Linnaeus, 1758 Infr aorder Catarrhini Geoffroy, 1812 Superfamily Cercopithecoidea Gray, 1821 Family Cercopithecoidae Gray, 1821 Subfamily Cercopithecinae Gray, 1821 Tr ibe Papionini Burnett, 1828 Genera Parapapio Hopwood, 1936 Species Parapapio ado Hopwood, 1936 The second specimen being examined is a nearly complete mandible from Locality 2 of the Upper Laetoli l beds. Absent from this specimen is the inferior base of the mandibular corpus, the right ascending ramus and the base of the left gonial angle. The inferior portion of the left as cending ramus is still intact (while still missing the inferior base of the mandibular corpus nd gonial angle); however, the coronoid process and the condylar process are absent. On the internal aspect of the left ascending ramus is the mandibular foramen with a superoposterior to inferoanterior displacement. Also present on the left interior ascending ramus is a partial lingula and a mylohyoid groove with a similar displacement as the mandibular foramen. The relative shape of the mandible is narrowly para bolic and has a relatively deep mandibular planum. The specimen contains a strong simian shelf (inferior transverse torus) that extends posteriorly to the level of P 4 M 1 The interior aspect of the simian shelf is a shallow depression inferior from the inc isors and premolars; however there is a deep digastric fossa placed anteriorly with a pitted depression on the internal side of the mandibular symphysis. On the external surface of the corpus is a well preserved mandibular symphysis that contains two groov es running superoanterior to the inferoposterior aspect, ending at the mental foramen on either side of the specimen.
33 The mandibular breadth at M 1 is 33.7 mm and the mandibular length at M 1 is 30mm giving a symphyseal ratio of 0.89 (symphyseal ratio = mand ibular length/mandibular breath). This indicates that the mandibular specimen has a lesser degree of symphyseal curvature than comparatively sized cercopithecines (Ravosa 1996). Colobines tend to have shorter mandibular lengths at a common mandibular bread th suggesting that they exhibit similar levels of positive allometry of mandibular length versus mandibular breadth. Since the mandibular length is shorter than the mandibular breadth, this suggests that this mandibular specimen from Laetoli may be colobin e. The intact dental arcade consists of complete right and left I 1 I 2 P 3 P 4 M 1 M 2 and M 3 The occlusal cusps are still intact for all dentition and there is relatively little sign of wear, except the left I 1 where the occlusal surface is partially m issing due to taphonomic processes not due to wear. However, the distal aspect of the left I 2 is missing a small portion of the enamel and dentine on the labial surface of the tooth itself. There is a diastema present between the canines and P 3 and presen t on P 3 is a cingulum due to the diastema formation. The buccal cusps from P 4 to M 3 which include the protoconid and hypoconid, are slightly worn compared to the lingual cusps a general conformity to a typical papionin (cercopithecine) wear pattern (Harrison & Harris 1994). Present on the distal aspect of the left M 3 is a fovea located posterior to the oblique line. The fossa accommodates for the 5 th distal cusp, the hypoconulid, on M 3 which is nearly erupted from its alveoli. Present on the lingua l aspect of the anterior teeth is a cingulum, located on the neck of the anterio r teeth. According to Harrison and Harris (1994) a cingulum is present in all extant African colobines; however, this feature is highly variably developed and is
34 better preserv ed on females. Although the cingulum is present on this mandibular specimen, it also supports the notion that this specimen is a colobine, though more diagnostic features are needed in order to distinguish the sex of the mandible once a species is determin ed The overall mandibular morphology of the specimen is similar to colobine mandibular morphology due to its relatively upright ascending ramus, a mandibular corpus with a relat ively constant height, and a si mian shelf that is relatively shallow mesially (Delson 1976). Although the gonial angle is absent, the expansion of this feature creates a deeper corpus (Delson 1976).
35 Figure 9 : M andibular specimen of Parapapio ado from Locality 2; occlusal view Figure 10 : M andibular specimen of Parapapio ado from Locality 2; lateral view
36 Figure 11 : Comparative mandibular specimens for the Parapapio ado Laetoli specimen from Locality 2; (Top: Papio Anubis; Middle: Colobus polykomos; Bottom: Parapa pio ado specimen from Locality 2
37 Methods: D ental M etrics Primate dentition: An introduction to the teeth of non human primates (2002), odontometric measurements were taken with digital dental calipers accurate to 0.01mm. The mesiodistal (MD), buccolingual (BL) measurements were also taken on each individual tooth. However, if certain measurements cannot be taken due to taphonomic processes, those od ontometrics were omitted in this research. More information on the measurement s taken are described in Table 2. The measurements of the righ t mandibular posterior teeth (P 3 M 3 ) of the unknown Laetoli speci mens are summarized in Table 3 whereas the summarized MD and BL measurements of the right mandibular posterior teeth of the ext ant and fossil prima tes are summarized in Appendix A Crown Shape (C shape ) and Crown Area (C area ) were also calculated using the MD and BL data (C shape = BL/MD 100; C area =MD*BL). The qualitative cusp terminology used in this discussion is taken from Delson (1976) and Szalay and Delson (1979).
38 Table 2 : Method of mandibular odontometrics taken for proposed Cerco pithecid specimens from Laetoli (Swindler 2002) Dental Type Dimensions taken Incisors MD: diameter at the incisal edge of the lower incisors BL: diameter taken at the cementoenamel junction at a right angle to the MD diameter Canines MD: (Lower Canines) MD diameter measured at the level of the mesial alveolar margin BL: diameter taken at the cementoenamel junction a t a right angle to the MD diameter Pre molars MD: Maximum MD diameter taken btween the contact point. If the mesial contact is lacking on P 3 or P 3 owing to a diastema between it and the canine, the maximum horizontal distance is measured from the distal c ontact point to the most mesial point on the surface of the premolar. Same method was utilized on primates with three premolars. BL: Maximum BL diameter taken at a right angle to the MD diameter Molars MD: Maximum MD diameter taken on the occlusal surface between the mesial and distal contact points BL: Maximum BL diameter measured at right angles to the MD dimension. The dimensions of both the tigonid and talonid were taken in this manner.
39 Figure 12 : Odontometric landmarks. B = buccolingual breadth; L = mesiodistal length. Darius (2002) Primate Dentition: An Introduction to the Teeth of Non Human Primates: Reprinted from Swindler, D.R. (1976) Dentition of Living Primates
40 Figure 13 : Identification of cercopithecid dental landmarks on the right lower M 3 of Theropithecus [modified after Jolly 1972] A = Occusal aspect; B = lingual aspect; C= buccal aspect; D= mesial aspect; (stippling indicates contact facet with M 2 ; E = distal aspect. a = Mesial buccal cleft; b = protoconid; c = median bucc al cleft; d = buccal margin; e = hypoconid; f = distal buccal cleft; g = hypoconulid; h = tuberculum sextum; I = distal fovea; j = hypolophid; k = entoconid; l = lingual margin; m = talonid basin; n = metaconid; o = metalophid; p = trigonid basin (mesial f ovea); q = mesial shelf; r = median lingual notch; s = distal lingual notch; t = distal buccal notch; u = median buccal notch. In this, elevated features (crests, ridges, outlines) are represented by solid lines, depressed features (groove, clefts) by dott ed lines. (Delson 1975 pp.176, Fig. 2)
41 Table 3 : Mandibular odontometrics from the proposed Cercopithecid sp. from Laetoli, Local 2 and LP 061703 01&02 fragment ; right mandibular segme nt from P 3 M 3 in mm TOOTH MD BL Unknown Cercopithecid sp from Laetoli LP 061703 01 & 02 Unknown Cercopithecid sp from Laetoli LP 061703 01 & 02 P 3 5.79 ------------5.62 ------------P 4 6.97 ------------6.67 ------------M 1 8.91 8.85 7.63 6.51 M 2 10.91 9.90 8.87 9.10 M 3 14.74 ------------9.68 ------------* MD Mesiodistal length, BL Buccolingual breadth (Maximum buccolingual breadth were used as the BL for all molars)
42 Methods: Statistical Analysis In order to interpret the specimen s from Laetoli, extant and fossil primates had to be considered in the analysis. The values taken from this dataset include the right portion of the mandible, so that it can be compared to the nearly complete mandible from Locality 2 and the L P mandibular fragment from Locality 1. This com parative sample consist mostly of e xtant primate d ata; however the Rhinocolobus fossil will be grouped with its extant population based on their morphological similarities they share, whereas the fossil Theropithecus data will be grouped as their own group Computational statistical analysis is performed in order to distinguish the recently recovered fossilized cercopith specimens found at Laetoli. A Prin ciple Component Analysis (PCA) and Discriminant Function Analysis (D FA) are used in this study. The PCA produces extraction factors, where the first extracted factor account s for the largest variance of the total variance inherent in the odontometric data. The eigenvalue also provides information on the equivalent number of variables the new factor r epresents. T he DFA will be used to distinguish whether or not odontometric data can be used to classify primate species by using a weighted combination of predictor values to classify a species into the predis posed criterion species groups. However, since the number within each species varies, a prior probability was specified in the DFA based on the proportionality of occurrence of the species in the total sample of cercopithecines and colobines. Also, the grouping variables used for the PCA and DFA consis ted of the genera that is associated with each species; therefore a total of 7 grouping variables were used ( Cercocebus, Cercopithecus, Colobus, Mandrillus, Papio, Theropithecus, and the unknown Laetoli specimens). High correct classification for the DFA m eans that there is
43 a high probability for the observed grouping variable to match the expected grouping variable; whereas the low correct classification means that there is a low probability for the observed grouping variable to match the expected grouping variable.
44 CHAPTER I V RESULTS P 3 The first two factors show distinct groupings between cercopithecids occupying forested habitats versus mosaic habitats However, the PC I has an eigenvalue of 3.06 and explains 76.66 % of the total variance, whereas PC II has an eigenvalue of 0.86 and explains for 21.61% of the total variance. Since PC I has the largest eigenvalue, the corresponding eigenvector exhibits the direction of the greatest variation (76.6 6%). The eig envectors associated with PC I shows that C shape is extracted, where MD, BL an d C area are retained variables since they are moderately correlated (r MD =0.56, r BL =0.49, r Carea =0.56). Figure 14 shows the graphical representation of the PCA for P 3
45 Figure 14 : 3D score scatterplot for the P 3 PCA In the DFA, describes the proportion of total variance in the discriminant scores not explained by the differences among groups. In this instance, the null hypothesis is that the means of MD, BL, and C area on the discriminant function at group centroids are equal. In Lamba ;p<0. 001), it indicates that MD, BL, and C area means differ. However, since the W
46 has more discriminatory ability. Given this, eigenvalues indicate the proportion of variance explained. The hi gh eigenvalue of 1.89 indicates that the discriminatory power of the function is strong a nd the high canonical correlation (r=0.81) indicates a strong positive association between discriminant scores and grouping which demonstrate a strong discriminating function for P 3 Of the 170 cases of MD, BL, and C area variables, 112 cases of th e variables were correctly cla ssified within their respective primate species, where 65.88 % of the predicted group membership matched the observed group membership of species. However, 58 cases of the variables were misclassified within the pred icted group membership, where 34.12 % of the predicted group membership did not match the observed group membership. P 4 The first two factors show less of a distinct grouping pattern, where forested cercopithecids and cercopithecids that exploit more mosaic environments are still separated in two groups with much more overlap between the loading factors. However, the PC I has an eigenvalue of 2.91 and explains 72.86 % of the total variance, whereas PC II has an eigenvalue of 1.07 and explains for 26.81% of the total variance. Since PC I has the largest eigenvalue, the corresponding eigenvector exhibits the direction of the greatest variation (72.86%). The eigenvectors associated with PC I shows that C shape is extracted which is also seen in P 3 where MD, BL and C area are retained variables since they are moderately correlated (r MD =0.57, r BL =0.57, r Carea =0.58). Figure 15 shows the graphical representation of the PCA for P 4
47 Figure 15 : 3D score scatterplot for the P 4 PCA Cercocebus amount of prediction to discriminant the groups in a DFA in the ;p<0.001), it indicates that MD, BL, and C area means differ. discriminatory abil ity. The high eigenvalue of 2.67 indicates that the discriminatory
48 power of the function is strong and the hi gh canonical correlation (r=0.85 ) indicates a strong positive association between discriminant scores and grouping which demonstrate a strong discriminating function for P 4 even t hough the grouping variables in the PCA were not clearly distinct. Of the 170 cases of MD, BL, and C area variables, 126 cases of the variables were correctly classified within their respective primate species, where 74.12 % of the predicted group membership matched the observed group membership of species. However, 4 4 cases of the variables were misclassified within the predicted group membership where 25.88 % of the predicted group membership did not match the observed gro up membership. M 1 Like P 4 the first two factors show less of a distinct grouping pattern, where forested cercopithecids and mosaic range cercopithecids are still separated in two groups with much less overlap between the loading factors but more of a spread between PC I and PC II. However, the PC I has an eigenvalue of 2 .95 and explains 73.77 % of the total variance, whereas PC II has an eigenvalue of 1.04 and explains for 25.99% of the total variance. Since PC I has the largest eigenvalue, the corres ponding eigenvector exhibits the direction of the greatest variation (73.77%). The eigenvectors associated with PC I shows that C shape is also extracted, where MD, BL and C area are retained variables since they are moderately correlated (r MD =0.57, r BL =0.5 8, r C area =0.58). Figure 16 shows the graphical representation of the PCA for M 1
49 Figure 16 : 3D score scatterplot for the M 1 PCA for the DFA it indicates that MD, BL, and C area near 1, it indicates that function 1 also has more discriminatory ability, similar to P 3 and P 4 The high eigenvalue of 4.60 indicates that the discriminatory power of the function is strong and the high canonical correlation (r=0.91) indicates a strong positive association between discriminant scores and grouping which demonstrate a strong discrimi nating function for M 1 even though the grouping variables in the PCA were not clearly distinct.
50 Of the 171 cases of MD, BL, and C area variables, 140 cases of the variables were correctly classified within their resp ective primate species, where 81 8 7% of t he predicted group membership matched the observed group membership of species. However, 31 cases of the variables were misclassified within the predicted group membership, where 18.13 % of the predicted group membership did not match the observed group mem bership. M 2 Like P 4 and M 1 the first two factors show less of a distinct grouping pattern, where forested and mosaic dwelling cercopithecids are still separated in two groups with some overlap between the loading factors but more of a spread between PC I and PC II. However, the PC I has an eigenvalue of 2.96 and explains 74 .10 % of the total variance, whereas PC II has an eigenvalue of 1.03 and explains for 25.63% of the total variance. Since PC I has the largest eigenvalue, the corresponding eigenvector exhibits the direction of the greatest variation (74.10%). The eigenvectors associated with PC I shows that C shape is also extracted (as seen in P 3 P 4 and M 1 ), where MD, BL and C area are retained variables since they are moderately correlated (r MD =0.58, r BL =0.57, r Carea =0.58). Figure 17 shows the graphical representation of the PCA for M 2
51 Figure 17 : 3D score scatterplot for the M 2 PCA Cercocebus the M 2 ;p<0.001), which indicates that MD, BL, and C area indicates that function 1 also has more discriminatory ability, similar to P 3 P 4 and M 1 The high eigenvalue of 6.07 indicates that the discriminatory power of the function is strong and the high canonical correlation (r=0.93) indicates a strong positive association between discriminant scores and grouping which demonstrate a strong discriminating function for M 1 even though the grouping variables in the PCA were not clearly distinct.
52 Of the 171 cases of MD, BL, and C area variables, 138 cases of the variables were correctly classified within their resp ective primate species, where 80.70 % of the predicted group membership matched the observed group me mbership of species. However, 33 cases of the variables were misclassified within the pre dict ed group membership, where 19.30 % of the predicted group membership did not match the observed group membership. M 3 Like P 3 the first two factors show a distinct grouping pattern, where forested and mosaic dwelling cercopithecids are still separated in two groups with no overlap between the loading factors with more of a variation between PC I and PC II. However, the PC I has an eigenvalue of 2.96 a nd explains 74.04 % of the total variance whereas PC II has an eigenvalue of 1.03 and explains for 25.67% of the total variance (which resembles M 2 PCA output). Since PC I has the largest eigenvalue, the corresponding eigenvector exhibits the direction of the greatest variation (74.04%). The eigenvectors associated with PC I shows that C shape is extra cted (which is similar to all tooth types) where MD, BL and C area are retained variables since they are moderately correlated (r MD = 0.58 r BL = 0.57 r Carea = 0.58 ). Figure 18 shows the graphical representation of the PCA for M 3
53 Figure 18 : 3D score scatterplot for the M 3 PCA In the DFA, the M 3 is the lowest of all the DFA performed and is b ;p<0.001), which indicates that MD, BL and C area means highly and is the lowest of the values compared to P 3 P 4 M 1 and M 2 it indicates that function 1 also has the most discriminatory ability in M 3 The high eigenvalue of 8. 17 indicates that the discriminatory power of the function is strong and the hi gh canonical correlation (r=0.94 ) indicates a
54 strong positive association between discriminant scores and grouping which demonstrate a strong discriminating function for M 3 Of the 170 cases of MD, BL, C shape and C area variables, 143 cases of the variables were correctly classified within their respect ive primate species, where 84.12 % of the predicted group membership matched the observed group me mbership of species. However, 27 cases of the variables were misclassified within the predict ed group membership, where 15.88 % of the predicted group membership did not match the observed group membership. Summary: Results The PCA output shows that only C shape is extracted from the data for all tooth types MD, BL and C area are retained variables for all tooth types since they are all moderately correlated to one another. MD, BL and C area are correlated because mesiodistal length and t he buccolingual width are rela ted to overall tooth mor phology Also, since C area =MD*BL we can assume that all three variables are significant to each other since calculation of C area includes both MD and BL variables Further analysis of the DFA will be discussed in the section below; however since C shape is not a significant variable (as shown in the PCA), the DFA will exclude the C shape variable into the analysis. The overall DFA shows that the buccal teeth ( M 1 M 2 and M 3 ) have a much higher discrimin atory power than the premolars. Here, the discriminato ry ability increases with M 1 and M 3 but M 2 has the lowest discriminatory power from all the molars yet still retains a higher discriminatory ability than the premolars.
55 CHAPTER V DISCUSSION Within Cercopithecidae are two sub families: Colobinae which are leaf eaters and Cercopithecinae which are cheek pouched monkeys (Delson 1975). Cercopithecines can then be subdivided into two tribes: Cercop ithecini and Papionini. As stated earlier, the prominent difference within Cercopithecidae is between th e longer faced cercopithecines and the shorter more upright faced colobines, although there is much overlap in proportio ns for the two (Delson 1975). According to Delson (1975) colobines are relatively wider in the fa ce, as seen in the wider interorbital region especially in the smaller species. However, there is some overlap between cercopithecines and colobines within the posterior aperture of the nasal cavity leading into the nasopharynx region (Delson 1975). The PCA results show a distinct grouping b etween forest and mosaic dwelling cercopithecids; such that the predominately forests cercopithecids include Colobus, Cercopithecus, and Cercocebus genera, and the mosaic dwelling cercopithecids include Mandrillus, Papio and Theropithecus. The two groups show dis tinctness since both grouping s show that cercopithecids can be divided by the type of environment they prefer and the overlap may be due to the mosaic enviro n ments preferred by papionins. Papionins are known to occupy a range of habitats such as tropical rainforest that usually borders a savannah, mountainous and plateau terrains, open grasslands with patches of dense thicket, and savanna woodlands.
56 Mandrillus, Papio and Theropithecus all follow a s imilar dental pattern, even though Theropithecus does possess specialized features from the other Papionin groups. The PCA may have grouped these genera together based on characteristics such as accessory cuspules, lower molar width relationships, and mode rate f laring and reduction of the mandibular incisor enamel (Delson 1975) indicating a more broad diet However, the greatest difference between Theropithecus and the other Papionins is its great crown height and heightened relief with deepened foveas. If the overall crown height were measured for each tooth type instead of the mesiodistal length and buccolingual width, Theropithecus would probably have its own classifica tion in the principal component analysis. The second grouping of the other cercopithecines and colobines exploit predominately forested environments, such as a riverine forest, swamp forests, moist evergreens, montane forests, and dense tickets found in savanna regions. This grouping also follows c ercopithecini dental patte rning which is characterized by the loss of M 3 hypoconulid and a reduction of M 3 which is also indicative of a predominately leaf eating diet. Other typical characteristics include elongate and low flaring teeth; however, there have been instances where cercopithecin e s have distinctive lateral flaring ( as seen in Cercocebus ) an increase of mesial width, and trigonid size from the mandibular M 1 to M 3 Cercopithecini also share the s ame general characteristics of p apionins (stated before); however, they do not possess access ory capsules and have unreduced lower incisor enamel as seen in colobines (Delson 1975). It is assumed that Cercopithecus would group closer with Colobus and Cercocebus since most Cercopithecus species are arboreal frugivores and some are considered to be leaf eaters (Szalay & Delson 1979).
57 S tudies have demonstrated that Cercocebus possesses a flatter crown surface of the molars suggesting a more frugivorous diet with teeth adapted to crushing rather than slicing (Szalay & Delson 1979). On the other hand, Kay (1978) found that unworn Cercocebus teeth plotted out closer to colobines and other folivorous Lophocebus species (Szalay & Delson 1979: Kay 1978). The unknown LP mandible groups slightly with the forest dwelling cercopithe cids (tooth types P 3 P 4 and M 2 ) in the PCA; however for the M 1 tooth type, the mandible is situated between both groups and its M 3 is situated in the mosaic dwelling cercopithecid group. On the other hand, s pecimen 061703 01&02 (fragmented M 1 and M 2 specimen) classifies with the mosaic dwelling cercopithecid group in the principle component analysis. We can see that the LP mandible possesses an intermediate classification between both groupings, as seen in the qualitative analysis; whereas s pecimen 0 61703 01&02 categorizes primarily with papionins indicating a preference of a more mosaic environment. The DFA for P3 shows that 35 Cercocebus measurements are correctly classified within the Cercocebus genus, whereas 13 are misclassified into Colobus and 3 are misclassified into Papio genera. Twenty five Colobus measurements were correctly classified into the Colobus genus, whereas 16 are misclassified into the Cercocebus 4 are misclassified into Papio and 1 is misclassified into Cercopithecus. 2 of the Mandrillus measurements are correctly classified, whereas 2 are misclassified into Papio and finally, 48 Papio measurements are correctly classified with 4 misclassified for both Colobus and Mandrillus and 3 misclassified in Theropithecus. On the other hand, the 2 Cercopithecus
58 and Theropithecus measurements were correctly classified. The LP mandible is misclassified into Theropithecus ; however, this is only for the P 3 tooth type. The DFA for P 4 shows that 41 Cercocebus measurements are correctly class ified, whereas 8 measurements are misclassified into Colobus 1 measurement is misclassified for both Papio and the LP mandible. Only 1 measurement of Cercopithecus is correctly classified and the other measurement is misclassified as Colobus. 33 Colobus measurements are correctly classified, whereas 8 measurements are misclassified into Cercocebus 4 measurements misclassified into Cercopithecus and 1 measurement misclassified into Papio. Like P 3 2 of the Mandrillus measurements are correctly classified, whereas 2 are misclassified into Papio Fourty nine Papio measurements are correctly classified, whereas 15 of the measurements are misclassified into Cercoebus. The only correctly classified measurement without any misclassifications is for Theropithecus The LP mandible is misclassified into Cercocebus ; however, this is only the case for the P 4 tooth type. The results of the DFA show s that the posterior teeth ( M 1 and M 3 specifically ) increase in discriminatory power and a have a lower percentage of misclassification than the premolars (decre ase in misclassification by 7.75 % from P 4 to M 1 ). The DFA for M 1 shows that 46 Cercocebus measurements are correctly classified, whereas only 5 are misclassified into Colobus; 33 of the Colobus measurements are correctly classified, whereas 9 measurements are misclassified into Cercocebus, 3 measurements misclassified into Cercopithecus and 1 measurement is misclassified into Papio, and finally 57 Papio measurements are cor rectly classified with 7 measurements
59 misclassified into Cercocebus. Measurements of Cercopithecus Mandrillus, and Theropithecus are correctly classified for the M 1 tooth type. Like P 4 the LP mandible is misclassified into Cercocebus whereas LP s pecimen 061703 01 is misclass ified into Papio. Although M 2 has a lower discriminatory power than M 1 and M 3 19.3% of the measurements were misclassified; however, M 2 has a lower percentage in misclassification when compared to P 4 (decrease in misclassification by 6.58%). Fourty three of the Cercocebus measurements are correctly classified, whereas 8 of the measurements are misclassified into Colobus ; 38 of the Cercocebus are correctly classified, whereas 3 measurements are misclassified into Cercocebus 4 measurements are misclassifie d into Cercopithecus and 1 measurement is misclassified as Papio ; and finally 57 Papio measurements are correctly classified with 6 measurements misclassified into Cercocebus. Like M 1 Cercopithecus, Mandrillus, and Theropithecus are correctly classif ied for the M 2 tooth type. Simi l a r to M 1 the LP mandible for M 2 is misclassified into Cercocebus whereas LP specimen 061703 02 is misclassified into Papio. A stated earlier M 3 has a high eigenvalue of 8.17, which indicates that the discriminatory power of the function is strong and the high canonical correlation (r=0.94) indicates a strong positive association between discriminant scores and grouping which demonstrate a strong discriminating function for M 3 M 3 has the lowest percentage of misclassification (15.88%) where 43 Cercocebus measurements are correctly classified with only 8 measurements misclassified as Colobus; 37 Colobus measurements correctly classified with 7 measurements misclassified as Cercoceb us and 1 measurement misclassified in Cercopithecus and Papio ; and finally 61 Papio measurements correctly
60 classified with 3 measurements misclassified as Cercocebus. Similar to M 1 and M 2 Cercopithecus, Mandrillus Theropithecus are all correctly classif ied for the M 3 tooth type. Unlike M 1 and M 2 the LP mandible misclassifies as Papio instead of Cercocebus which is similar to the LP specimen 061703 01& 02 misclassification Although the unknown LP specimens were misclassified into different subtribes for all tooth types, they can all be classified into Papionini indicating that the tooth type for both specimens are similar to papionins. However, this contradicts the mandibula r morphology seen in the LP mandible since it resembles colobines instead of cercopithecines (which papionins are taxonomically classified under cercopithecines). The conclusion is that the LP mandibular specimen probably depended on a similar diet as cerc opithecines rather than a typical leaf eating diet o f colobines. Studies have shown that colobines and gibbons share many ancestral cranial and craniofacial patterns, while Cercopithecinae are the most derived and specialized group (Delson 1975, Vogel 196 6) However, it is assumed that early cercopithecid facial skeleton may have been muc h more similar to colobines than present day cercopithecines (Delson 1975, Vogel 1966). The LP mandible is a clear example of having derived tooth morphologies, as seen in papionins; yet still retains a colobine facial s keleton, a characteristic of early cercopithecids. The unknown LP ma ndible has a symphyseal ratio of 0.89, which indicates that the mandibular specimen has a lesser degree of symphyseal curvature than comp aratively sized cercopithecines (Ravosa 1996). As stated earlier, c olobines tend to have shorter mandible lengths at a common mandibular breadth suggesting that they exhibit similar levels of positive allometry of mandibular length versus mandibular breadt h. Since the
61 mandibular length is shorter than the mandibular breadth, this suggests that this mandibular specim en from Laetoli retains colobine ancestral features wit h more derived tooth morphology during the Pliocene epoch. According to Jolly (1966), the original adaptive niche of the cercopithecids was leaf eating and that the molar pattern is the most primitive character state. However, other studies have shown that colobines are morphologically ancestral in all character states since they are arbore al and could not be descended from terrestrial cercopithcine s, which is more indicative of environmental functionality than ancestry (Delson 1975, Napier 1970). The most appropriate understanding of cercopithecid evolution comes from the work of Vogel (196 6), wher e colobine and gibbon cranium are probably representative of an ancestral condition for ca tarrhines (Delson 1975). Modern colobines are able to process large amounts of foliage through an enlarged and sacculated stomach and modified digestive tract However, Delson (1975) suggests that although this characteristic may seem derived in nature, it may actually be an ancestra l trait for Colobinae. M odern cercopithecines possess buccal pouches for temporary storage of food prior to mastication, but acqui red a digestive system similar to other hominoids (Delson 1975) that may relatively affect tooth morphology. Another aspect of catarrhine evolution that may be useful in speciation is the number of chromosomes in a eukaryotic cell (Delson 1975) Gibbons and some colobines share the same number of diploid chromosomes ( 2n=44) which Delson suggests may be the ancestral number fo r all catarrhines. On the other hand, papionins have a diploid number of 42, indicating that papionins
62 However, species of Cercopithecini have a diploid number of chromosomes that range from 48 to 72, which indicates the strongest specialization according to Delson (1975). Napier (1970) suggest that Old World monkeys originated as an adaptive unit where there was a selective advantage for primates that were able to subsist on leaves rather than fruits when necessary, implying that forests and the surrounding environment were seasonal. However, the ancestors o f cercopithecids may not have been dependent on leaves, but were able to supplements the basic primate diet of fruit with foliage in habitats where they were abundant (or seasonal changes where foliage were abundant during that time of year). Considering N cercopithecid evolution (Napier 1970, Delson 1975, Jolly 1966) early cerecopithecids probably had a colobine like cranium with macaque like teeth (Tribe: Papionini) and long limbs. Delson (1975) suggests that the ea rly cercopithecids were arboreal quadrupeds who ate fruits when possible but supplemented their diet with leaves under certain ecological conditions. By the Middle Miocene, colobine ancestors shifted to a more folivorous diet with associated changes to te eth morphology and more specialized gut; whereas cercopithecine ancestors started to shift towards a omnivorous diet leading into the specialization of the buccal pouches and allometrically longer faces. The unknown Laetoli specimens follow the latter traj ectory, where the colobine like mandibular morphology retains the primitive state, yet develops specialization for a more omnivorous diet that are seen in early cercopithecids.
63 The DFA statistical analysis for the unknown Laetoli specimens classify into Theropithecus, Cercocebus and Papio all derived from the same p apionini tribe Although the mandibular morphology resembles extant colobines, the tee th morphology is indicative of p apionini, which is a quality of an early cercopithecid. Knowing that the unknown species cla ssifies under p apionini, this narrows down which species these specimens are By utilizing the statistical analysis and qualitative descriptions, the unknown Laetoli specimens can be attributed to Parapapio ado (Pp. ado) which is also t he most common cercopithecid found at Laetoli. Parapapio ado also posse s ses a relatively short face during its subadult phase (Harrison 2011), which is recognizable through the overall shape and breadth of the LP mandibular specimen. The LP mandibular spe cimen is a subadult female due to the mid erupted M 3 and reduced canine size and morphologies seen in P 3 which is also a sexually dimorphic tooth However, the LP specimen 061703 01&02 cannot be classified as a male or female due to the lack of qualitative diagnostic features. A more comprehensive analysis must be completed in order to classify the sex of the specimens; however based on the qualitative and statistical analysis, the LP mandible implies subadult female characteristics. The recover ed LP mandible exhibits similarities to Parapapio ado by the presence of 3 mental foramen; however most species (and is present on the LP mandible) of Parapapio ado have a common mental foramen found on the inferior two thirds of the mandibular corpus located below P 4 and M 1 which is positioned anterolaterally (Harrison 2011). Like Parapapio ado, the P 3 of the LP mandibular specimen is short and relatively narrow wit h a slight crown extension mesiobuccally. Its P 3 also has a distinctive
64 protoconid with a mesial crest that extends to the lingual cingulum. As stated earlier, the P 4 of the L P mandible possesses a pyramida l protoconid; however, the metaconid is unrecogniz able and may be worn due taphonomic processes or attrition. On the other hand M 1 the buccal cusps (protoconid and hypoconid) are more conical and are seen slightly more elevated than the lingual cusps (metaconid and entoconid); however, this height differ ential will decrease with the amount of wear (Harrison 2011). Although the protoconid and hypoconid are usually seen equal in size for Parapapio ado the LP mandible shows a very slight increase in protoconid height. M 2 has similar characteristics to M 1 ; however it is clear that the LP mandible possesses a protoconid and hypoconid that are subequal in size. Like Parapapio ado the M 3 of the LP specimen possesses a hypoconulid that is positioned relatively low on the crown on the distal aspect, where the entoconid is situated closer to the hypoconid than the protoconid to the metaconid. Also present on the M 3 is a triangular distal fovea with a distolingual tuberculum sextum on the occlusal surface, which is also variably present on other Parapapio ado spe cimens (Harrison 2011). As stated earlier, the LP specimens classify under Theropithecis, Cercocebus and Papio. However, the Cercocebus and Papio material are from extant species, whereas the Theropithecus used in the analysis is from the fossil Theropi thecine material from Southern Angola (Jablonski 1994) and Laetoli (Leakey 1969) Parapapio ado has many similarities with extant Cercocebus (Lophocebus) and Papio such as a broader and shorter P 3 crown relative P 4 size that is intermediary between Cercocebus Mandrillus and Lophocebus Papio Theropithecus and share basic configuration of molar
65 morphologies (Freedman 1957; Szalay and Delson 1979; Leakey and Delson 1987; Frost and Delson 2002).
66 CHAPTER VI CONCLUSION Parapapio ado and other fossil cercopithecids, such as Galago sadimanensis and Paracolobus sp., is suggestive of a close d woodland environment or forests at Laetoli. However, extant c ercopithecids tend to thrive in a range of habit ats from grasslands to forests, where s tands of trees and rocky outcrops are common sleeping quadrants. Given this, the newly recovered specimens that were able to be quantifiable can be attributed to Parapapio ado This particular species is the most common cercopithecid found at Laetoli during the Pliocene epoch which also supports the existing literature that the paleoenvironment of Laetoli consisted of a denser forested environment than previously thought. Previ ous analysis of odontometric data (Delson 1975; Vogel 1966; Jablonski 1994; Frost 2001) has been shown to compare fossil and extant cercopithecids; however, establishing speciation is usually based on qualitative data. Although three of the five specimens were unable to be classified into a particular species due to the high fragmentation and poor preservation, two specimens were able to be measured and quantified. This research has demonstrated that quantitative data can be utilized in deciphering fossil and extant species of cercopithecids since tooth morphology is representative of dietary adaptations and environmental contexts. The principle component analysis de monstrated that odontometric data can be used in general groupings of speciati on based on di etary adaptations and habitat preferences ; however C shape was irrelevant to the analysis and was extracted. MD, BL and C area were all retained
67 variables that were h ighly correlated and could be used for a better approximation of speciation for the discriminant functional analysis. The DFA showed that the posterior buccal teeth had a higher discriminatory power than the premolars; however M 3 had the highest discriminatory ability in deciphering primate species and exhibited the lowest miscla ssification percentage for all tooth types. The Parapapio ado specimens have shed light on the evolutionary context of cercopithecids during the Pliocene at Laetoli, which is also supported by the ancestral colobine gibbon like cranium and derived papionin tooth morphology. The characteristics of these specimens suggest and support that cercopithecids during the Pliocene may have had a combination of ancest ral and derived characteristics that helped them evolve towards a more omnivorous diet
68 REFERENCES Bird DW & ( 2006 ) Behavioral Ecology and Archaeology. J Archaeol Res 14:143 188. Deino AL ( 2011 ) Paleontology and Geology of Laetoli: Human Evolution in Context. Volume 1: Geology, Geochronology,Paleoecology and Paleoenvironment Vertebrate Chapter 4:40Ar/39Ar Dating of Laetoli, Springer .77 97. Delson E (1975) Evolutionary history of the Cercopithecidae. In F.S. Szalay (Ed.) Approaches to primate paleobiology. Contributions to primatology (Vol. 5, pp.167 217. Basel: Karger. Foley R (2005) The adaptive legacy of human evolution: a search for the environment of evolutionary adaptedness. Evo l utionary Anthropol ogy 194 203. Freedman L (1957) The fossil Cercopithecoidea of South Africa. Annals of the Transvaal Museum. 23:121 262. Frost SR (2001) New Early Pliocene Cercopithecidae (Mammalia: Primates) from Aramis, Middle Awas h Valley, Ethiop ia. American Museum of Natural History. 3350: 1 36. Frost S (2001) Fossil Cercopithecidae of the Afar Depression, Ethiopia: Species Systematics and Comparison to the Turkana Basin. PhD Dissertation from the City University of New York. Frost SR & Delson E (2002) Fossil Cercopithecidae from the Hadar formation and surrounding areas of the Afar Depression, Ethiopia. Journal of Human Evolution. 43:687 748. Groves, C. P. (2005). Wilson, D. E.; Reeder, D. M, eds. Mammal Species of the World (3rd ed .). Baltimore: Johns Hopkins University Press. p. 165. OCLC 62265494. ISBN 0 801 88221 4. Harris JM (1978) Palaeontoloy. In MG Leakey and RE Leakey (Eds.): Koobi For a Research Project. Col. I, The Fossil Hominids and an Introduction to Their Context. Ox ford. Clarendon Press. 32 53. Harrison T (2011) Paleontology and Geology of Laetoli: Human Evolution Volume I : Geology, Geochronology, Paleoecology and Paleoenvironment. Chapter 3 : Sedimentology, Lithostratigraphy and Depositional History of the Laetoli Area Springer Science+Business Media B.V.
69 Harrison T (2011) Paleontology and Geolog y of Laetoli: Human Evolution. In Volume II: Fossil Hominins and the Associated Fauna, Vertebrate Paleobiology and Paleoanthropology Chapter 6: Cercopith ecids Cercopithecidae, Primates. Springer Science+Business Media B.V. Harrison T & Harris EE (1995) Plio Pleistocene cercopiths from Kanam East, western Kenya. J ournal of Hum an Evol ution 30:539 561. Hay RL (1987) Geology of the Laetoli area. In: Leakey MD, Harr is JM (Eds.), Laetoli: A Pliocene Site in Northern Tanzania Clarendon Press, Oxford, pp. 23 47. Jablonski N (1994) New Fossil Cercopithecid Remains From the Humpata Plateau, Southern Angola. American Journal of Physical Anthropology 94:435 464. Kay RF (1978) Molar Structure and Diet in xtant Cercopithecidae. In: Studies in the Development, Function, and Evolution of Teeth (PM Butler and K Joysey, eds.) pp.309 339. Academic Press, London. Leakey MG (1982) Extinct Large Colobines From the Plio Pleistocene of Africa. Am erican J ournal of Phys ical Anth ropology 58:153 172. Leakey MD & Hay RL (1979) Pliocene footprints in the Laetolil Beds at Laetoli, northern Tanzania. Nature 278, 317 323. Leakey MG (1982) Extinct Large Colobines From th e Plio Pleistocene of Africa. American J ournal of Physical Anthropology 58:153 172. Leakey MD, Hay RL, Curtis GH, Drake RE, Jackes MK, & White T. (1976) Fossil hominids fro m the Laetoli Beds. Nature 262: 460 466. Leakey MG & Delson E (1987) Fossil Cercop ithecids from the Laetoli Beds. In MD Leakey & JM Harris (Eds), Laetoli: A Pliocene site in northern Tanzania (pp.91 107). Oxford: Clarendon Press Musiba C (1999) Laetoli Pliocene paleoecology: a reanalysis via morphological and behavioral a pproaches. Ph.D. Dissertation, University of Chicago, Chicago, Illinois. Napier JR (1970) Paleoecology and catarrhine evolution. In Napier and Napier (Eds), Old World Monkeys Academic Press, New York. 55 95. Ravosa MJ (1996) Jaw Morpology and Function in iving and Fossil Old World Monkeys. International Journal of Primatology 17: 909 932. Reed KE (1997) Early hominin evolution and ecological change through the African Plio Pleistocene Journal of Human Evolution 32: 289 322.
70 Strasser E and Delson E (1987) Clad istic analysis of cercopithecid relationships. Journal of Human Evolution 16:81 99. Su DF & Harrison T (2007) The paleoecology of the Upper Laetolil Beds at Laetoli: A reconsideration of the large mammal evidence. In Bobe Z, Alemseged Z, & Behrensmeyer (E ds), Hominin Environments in the East African Pliocene: An Assessment of the Faunal Evidence Springer. 279 313. Su DF & Harrison T (2008) Ecological implications of the relative rarity of fossil hominins at Laetoli. Journal of Human Evolution 55:672 681 Swindler DR (2002) Primate dentition: An introduction to the teeth of non human primates. Cambridge: Cambridge University Press. Vogel C (1966) Morphologische Studien am Gesichtsschadel catarrhiner Pimaten. Bibl. Primat., No. 4, pp. 1 226 (Karger, Basel 1966) Walker AC (1987) Fossil Galaginae from Laetoli. In: Leakey MD, Harris JM (Eds.), Laetoli: A Pliocene Site in Northern Tanzan ia. Clarendon Press, Oxford. 88 91. Wesselman HB (1984) The Omo micromammals: Systematics and paleoecology of early ma n site from Ethiopia (Contributons to vertebrate evolution, Vol. 7). Basel: Krager. Wolfheim, JH (1983) Primates of the World: Distribution, Abundance and Conservation. New York Zoological Society, University of Washington Press. Seattle and London.
71 APPENDIX A. Summary of Collected Cercopithecine and Colobine Odontometrics
72 B. Summary of E igenvalues from the PCA for P 3 M 3 Principle Component I Principle Component II P3 3.06 0.86 P4 2.91 1.07 M1 2.95 1.04 M2 2.97 1.03 M3 2.96 1.03 C. Summary of Percentage of T otal v =V ariance from the PCA for P 3 M 3 Principle Component I Principle Component II P3 76.66 21.61 P4 72.86 26.81 M1 73.77 25.99 M2 74.10 25.63 M3 74.04 25.67 D. Summary of the DFA R esults for P 3 M 3 Tooth Type N (p<0.001) Eigenvalue Canon Correlation (r= ) P3 170 0.22 1.89 0.81 P4 170 0.18 2.67 0.85 M1 171 0.11 4.60 0.91 M2 171 0.10 6.07 0.93 M3 170 0.07 8.17 0.94 E. Summary of the Correctly Classified and Misclassified P ercentages from the DFA for P 3 M 3 Tooth Type Percent (%) Correctly Classified Percent (%) Misclassified P3 65.88 34.12 P4 74.12 25.88 M1 81.87 18.13 M2 80.70 19.30 M3 84.12 15.88