New sivaladapid primate from Lower Siwalik deposits surrounding Ramnagar (Jammu and Kashmir State), India

Journal of Human Evolution 102 (2017) 21e41

Contents lists available at ScienceDirect

Journal of Human Evolution journal homepage: www.elsevier.com/locate/jhevol

New sivaladapid primate from Lower Siwalik deposits surrounding Ramnagar (Jammu and Kashmir State), India Christopher C. Gilbert a, b, c, *, Biren A. Patel d, e, N. Premjit Singh f, Christopher J. Campisano g, h, John G. Fleagle i, Kathleen L. Rust a, Rajeev Patnaik f a

Department of Anthropology, Hunter College of the City University of New York, 695 Park Avenue, New York, NY 10065, USA PhD Programs in Anthropology and Biology, Graduate Center of the City University of New York, 365 Fifth Avenue, NY 10016, USA New York Consortium in Evolutionary Primatology, New York, NY, USA d Department of Cell and Neurobiology, Keck School of Medicine, University of Southern California, Los Angeles, CA 90033, USA e Human and Evolutionary Biology Section, Department of Biological Sciences, University of Southern California, Los Angeles, CA 90089, USA f Department of Geology, Panjab University, Chandigarh, 160 014, India g School of Human Evolution and Social Change, Arizona State University, Tempe, AZ 85287, USA h Institute of Human Origins, Arizona State University, Tempe, AZ 85287, USA i Department of Anatomical Sciences, Stony Brook University, Stony Brook, NY 11794, USA b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 16 March 2016 Accepted 11 October 2016

Over the past century, numerous vertebrate fossils collected near the town of Ramnagar, India, have proven to be important for understanding the evolution and biogeography of many mammalian groups. Primates from Ramnagar, though rare, include a number of hominoid specimens attributable to Sivapithecus, as well as a single published mandibular fragment preserving the P4-M1 of the Miocene adapoid Sivaladapis palaeindicus. Since 2010, we have renewed fossil prospecting in the Lower Siwalik deposits near Ramnagar in an attempt to better understand the evolution, biogeographic timing, and paleoclimatic context of mammalian radiations in Asia, with a particular focus on primates. Our explorations have resulted in the identification of new fossil localities, including the site of Sunetar. The age of Sunetar and the Ramnagar region, in general, is tentatively dated between 14 and 11 Ma. In 2014, a partial right mandible of a sivaladapid primate was recovered at Sunetar, preserving the corpus with P4 roots and worn M1-M3 dentition. Although sivaladapids are known by numerous specimens of two genera (Sivaladapis and Indraloris) at Lower Siwalik sites on the Potwar Plateau (Pakistan) and at the Middle Siwalik locality of Haritalyangar (India), this new specimen is just the second sivaladapid recovered from the Ramnagar region. Our analyses suggest that the new specimen is distinct from all other sivaladapids, and we therefore describe it as a new genus and species close to the base of the Sivaladapinae. © 2016 Elsevier Ltd. All rights reserved.

Keywords: Ramadapis sahnii Sivaladapis Indraloris Chinji Sunetar Miocene

1. Introduction Fossiliferous deposits in the Siwalik Hills surrounding the town of Ramnagar (Jammu and Kashmir State), India, have been known to paleontologists for almost a century (Brown et al., 1924; Pilgrim, 1927; Vasishat et al., 1978). In fact, the earliest published records documenting systematic collection of material from the area derive from the Barnum Brown American Museum of Natural History Expedition, 1921e1923. Brown was tipped off to the potential of the Ramnagar area in 1922 by Charles Middlemiss, a prominent

* Corresponding author. E-mail address: [email protected] (C.C. Gilbert). http://dx.doi.org/10.1016/j.jhevol.2016.10.001 0047-2484/© 2016 Elsevier Ltd. All rights reserved.

geologist and Superintendent of the Mineral Survey in the state of Jammu and Kashmir at the time (Brown, 1922; Fermor, 1945). Middlemiss also collected in the area himself, recovering a hominoid partial mandible later described by Pilgrim (1927) as Sivapithecus middlemissi in his honor. How and when Middlemiss came to know of the fossiliferous deposits surrounding Ramnagar is not published, but since he worked extensively in Jammu and Kashmir beginning in 1908 (Fermor, 1945), and was Superintendent beginning in 1917, it is likely that he was generally aware of the area before 1922, and it is also possible that other paleontologists (amateur or otherwise) collected there even earlier. Brown published the first primate from Ramnagar, a hominoid partial jaw named Dryopithecus pilgrimi (Brown et al., 1924), now recognized (along with all other hominoid specimens from the

22

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

Ramnagar region) as Sivapithecus indicus (Kelley, 2002, 2005). Brown et al. (1924), as well as later authors (e.g., Pilgrim, 1927; Colbert, 1935; Gregory et al., 1938), suggested the area and its fauna correlated well with the better-known Lower Siwaliks Chinji Formation fauna in the Salt Range region of the Indian Potwar Plateau, now in Pakistan. Collection in the Ramnagar area has persisted on and off since Brown's expedition, with notable faunal studies by Vasishat et al. (1978), Gaur and Chopra (1983), Nanda and Sehgal (1993), and Basu (2004) confirming a correlation with the classic Chinji fauna in Pakistan and again indicating a Lower Siwalik age for the area. More recent studies have reported the presence of rodent taxa possibly suggesting a more specific Lower Chinji age for the Ramnagar deposits, ~12.7e14 Ma (Parmar and Prasad, 2006; Sehgal and Patnaik, 2012; Patnaik, 2013; Gilbert et al., 2014; Parmar et al., 2016). However, these estimates remain to be confirmed by more complete specimens and a better understanding of Ramnagar chronostratigraphy. Since the early 20th century (e.g., Brown et al., 1924; Pilgrim, 1927; Gregory et al., 1938), numerous additional hominoid specimens from the Ramnagar region have been recovered, mostly consisting of isolated teeth (Kelley, 2002, 2005). Thomas and Verma (1979) reported on the only other primate found in the area, a partial lower jaw preserving P4-M1 of the sivaladapid adapoid Sivaladapis palaeindicus, discovered at a locality ~2.5 km southeast of Ramnagar (Fig. 1). Since then, no new fossils definitively expanding the known primate taxonomic diversity at Ramnagar beyond S. indicus and S. palaeindicus have been found. In 2010, we renewed fossil prospecting and collecting in the Lower Siwalik deposits surrounding Ramnagar in an attempt to better understand the evolution, biogeography, timing, and paleoclimatic context of mammalian radiations in Asia, with a particular focus on primates and the chronology of the Ramnagar region. Between 2010 and 2015, we conducted six field seasons in the area, each approximately 10e15 days in length. To date, our explorations have resulted in the identification of new fossil localities in the Ramnagar area (Fig. 1; see also Gilbert et al., 2014) and the extension and clarification of the existing stratigraphic framework of the area (e.g., Basu, 2004). In October 2014, a partial

mandible of a sivaladapid primate was recovered by one of the authors (NPS) at the site of Sunetar (Fig. 1). This specimen, VPL/ RSP1 (Vertebrate Paleontology Laboratory, Panjab University Department of Geology/Ramnagar Sunetar Primate 1), preserves the right mandibular corpus under P4-M3 with P4 roots and worn M1-M3 dentition. Although sivaladapids are known by numerous specimens of at least two genera (Sivaladapis and Indraloris) at Lower Siwalik sites on the Potwar Plateau and at the Middle Siwalik locality of Haritalyangar, this new specimen is just the second known sivaladapid primate from the Ramnagar region. Based on comparisons with other known sivaladapid taxa, we describe this specimen here as a new genus and species. Our analyses suggest that this new sivaladapid taxon may lie near the ancestry of the other two Siwalik genera, Indraloris and Sivaladapis, and thus provide evidence of a diverse sivaladapine clade in the Mid-Late Miocene of South Asia. 1.1. Geological context The Siwalik Group of rocks, almost 7000 m in thickness, are exposed along the southern limb of the Suruin-Mastgarh anticline in the Mansar-Uttarbani section, Jammu District (Gupta and Verma, 1988). These deposits comprise the Mansar (Lower Siwalik), Dewal and Mohargarh (Middle Siwalik), and Uttarbaini and Dughor (Upper Siwalik) Formations (Fig. 1). The Mansar Formation is further divided into the lower Dodenal and the upper Ramnagar Members (Gupta, 1997, 2000). The Dodenal Member is characterized by the presence of thick massive sandstones, whereas the Ramnagar Member preserves thick mudstones and paleosols alternating with intraformational sandstones. The Ramnagar Member sequence that forms a part of the southern limb of the Udhampur Syncline can be best seen around the town of Ramnagar, situated approximately 38 km northeast of Jammu. Basu (2004) recognized 10 reference sandstones (AeJ, youngest to oldest) in the upper 350 m of the Ramnagar Member to provide a temporal-spatial context for almost all of the fossil localities in the area. The Sunetar site complex lies in the vicinity of where Basu (2004) marked the location of reference sandstone H (just below midway in the sequence) and we had

Figure 1. Left: General geological map of the Siwalik Group surrounding Ramnagar. Right: Close-up satellite imagery (GeoEye-1) of the Ramnagar region (corresponding to the dashed zone in insert of the geological map). Specimen VPL/RSP1 comes from the locality Sunetar 2. The previous sivaladapid specimen was reported to have been found ~2.5 km SE of Ramnagar, but no additional information was given (Thomas and Verma, 1979).

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

tentatively placed it just below this reference sandstone (Gilbert et al., 2014). However, in our recent survey, we realized that the correlation between, and extrapolation of, sandstones is likely too simplistic, and further geological survey is required to place this site in the overall stratigraphic framework of the Ramnagar sequence. The sivaladapid specimen was recovered on the surface at a locality referred to as Sunetar 2, which appears to be a part of the same stratigraphic package as Sunetar 1, a fossiliferous exposure ~1.2 km northwest of Sunetar 2 (Fig. 1) previously mentioned by Gilbert et al. (2014). Together, these two exposures are referred to as the site complex of Sunetar in this paper. GPS coordinates of the approximate site of discovery at Sunetar 2 are 32
45.9850 N, 75
19.4820 E at an elevation of ~1150 m. The exposure at Sunetar 2 is ~26 m thick and, like the rest of the Ramnagar sequence, is characterized by reddish brown mudstones, paleosols, and occasional thin (<1 m), very fine-grained sands sandwiched between two massive sandstones. There is a pseudoconglomerate layer composed primarily of mud pebbles, nodules, and concretions that becomes more concretionary at Sunetar 1, where it is full of coprolites and micro- and macro-vertebrate remains. The repetitive sequence of thick sandstones overlain by

23

Table 1 Partial faunal list for the site of Sunetar through 2015. Sunetar PISCES Indet. REPTILIA Testudines Indet. Crocodylia Indet. Squamata Acrochordidae Acrochordus dehmi Booidea Indet. cf. Colubroidea Indet. MAMMALIA Primates Sivaladapidae Gen. et sp. nov. Proboscidea Gomphotheriidae Indet. Perissodactyla Rhinocerotidae Indet. Artiodactyla Suidae Conohyus cf. sindiensis Listriodon cf. pentapotamiae Tragulidae Dorcatherium sp. Small Dorcatherium sp. Large

reddish brown mudstones and paleosols is pervasive throughout the Ramnagar sequence and represents major fluvial channels (sandstones) and overbank/floodplain deposits (mudstones/paleosols). The new sivaladapid specimen was collected atop an ~5 m thick massive sandstone at the base of the Sunetar 2 exposure, but was derived from higher up in the sequence based on its preservation and lack of sand matrix (Fig. 2). A partial faunal list for Sunetar is provided in Table 1. Overall, as has been noted for much of the past century, the fauna at Sunetar and the Ramnagar region more generally are consistent with a Chinji-level age, ~14e11 Ma (see also Gilbert et al., 2014). Ongoing field and lab work aimed at the collection of time-sensitive micromammals hopes to further refine the biostratigraphy and chronology at Sunetar, and more broadly the Ramnagar region. 2. Materials and methods 2.1. Comparative sample and imaging

Figure 2. Simplified stratigraphic section and photo of sequence at Sunetar 2. VPL/ RSP1 was found on the surface of a major sandstone (similar to where the gray GSP unit in the photo sits) but was derived from an overlying unit.

Fossil specimens examined in this study (Tables 2 and 3) derive from collections at the American Museum of Natural History (AMNH); the Yale Peabody Museum (YPM); the Geological Survey of Pakistan (GSP); the Geological Survey of India (GSI); the Institute of Vertebrate Paleontology and Paleoanthropology, Beijing (IVPP); the Paleontological Museum, University of Uppsala (PMU); the Department of Mineral Resources, Bangkok (TF); Panjab University Department of Anthropology (PU); and the Vertebrate Paleontology Laboratory, Panjab University Department of Geology (VPL). Where available, observations and measurements were taken on original specimens using digital calipers. Additional comparative observations and measurements were taken from high-quality casts, published photographs, and data provided in the literature. High-resolution micro-CT scans of VPL/RSP1 were obtained to facilitate anatomical description. The specimen was scanned at the

Taxon

Specimen number

24

Table 2 Comparative absolute measurements of the mandible and lower dentition for VPL/RSP1 and other sivaladapid primates.a P4 MD

P4 BL

M1 MD

M1 MDTri

M1 BL

M2 MD

M2 MDTri

M2 BL

M3 MD

VPL/RSP1

(5.6)

e

4.73

1.38

3.42

4.80

1.74

3.78

5.03

Sivaladapis palaeindicus

YGSP 25441

6.94

4.58

e

e

e

e

e

e

e

Sivaladapis palaeindicus Sivaladapis palaeindicus

GSI De224 YPM 19134 BSPHG 1939 X5 YGSP 31881 YGSP 32150 YGSP 32154 YGSP 46458 YGSP 31720 YGSP Se394 AMNH 103370

6.90 e

4.80 e

e e

e e

e e

5.91 e

2.35 e

4.61 e

7.11 7.02

Sivaladapis palaeindicus Sivaladapis Sivaladapis Sivaladapis Sivaladapis Sivaladapis Sivaladapis

palaeindicus palaeindicus palaeindicus palaeindicus palaeindicus palaeindicus

Sivaladapis palaeindicus

1.79 e 2.19 2.08

8.57

8.40

10.10

e

e

e

e

5.02 5.09

e e

e 12.76

e 13.24

This study Flynn and Morgan (2005) This study This study Flynn and Morgan (2005) This study This study This study This study This study This study

e

e

e

7.30

e

5.25

e

e

e

e

e

e e 6.54 5.76 e e

e e 2.41 2.38 e e

e e 5.16 4.50 e e

e e e 6.02 e e

e e e 2.29 e e

e e e 4.66 e e

7.35 7.39 e e e e

2.21 2.29 e e e e

5.21 4.70 e e e e

e e e 10.48 11.60 e

e e e 9.60 10.58 e

e e e e

e

e

e

e

e

e

e

e

7.26

2.21

4.71

e

e

e

This study Thomas and Verma (1979)

e

6.40

e

4.70

e

e

e

e

Sivaladapis palaeindicus

Average

6.88

4.59

6.23

2.40

4.79

6.41

2.32

4.84

7.23

LUVP 14500/ 116860 LUVP 14503 GSI 18093 (TYPE)

6.03

4.30

6.08

2.22

4.20

6.14

1.84

4.68

6.39

5.93

4.40

e

e

e

e

e

e

e

e

e

e

5.36

1.89

4.36

5.91

1.90

4.84

e

e

PUA 826e69

3.59

Reference

e

4.40

Sivaladapis nagrii

Mand. Ht. at M3

e e e e e e

6.80

Sivaladapis nagrii

Mand. Ht. at M2

e

Unnumbered

Sivaladapis nagrii

Mand. Ht. at M1

e e e e e e

Sivaladapis palaeindicus

Sivaladapis nagrii

M3 BL

6.00

4.80

5.90

e

4.70

(7.0)

e

5.00

(7.2)

e

e

e

e

e

e

2.20

4.95

11.04

10.98

13.24

1.56

4.56

e

e

e

This study

e

e

e

e

This study

e

e

e

e

e

(4.8)

e

e

e

e

Sivaladapis nagrii

PUA 343e69

e

e

e

e

e

6.80

e

5.50

6.90

5.00

e

e

e

Sivaladapis nagrii Sivaladapis nagrii Sivaladapis nagrii Sivaladapis nagrii

PUA 72e10 PUA 736e69 Averageb Rangeb

e 6.35 6.09 5.90e 6.35

e 4.42 4.60 4.40e 4.80

5.40 5.72 5.66 5.40e 5.90

e (1.96) 2.02 1.89e 2.22

4.60 4.53 4.63 4.30e 4.70

6.20 e 6.31 5.90e 7.00

e e 1.87 1.84e 1.90

5.00 e 5.01 4.90e 5.50

e e 6.93 6.70e 7.20

e e 1.56 1.56

e e 4.75 4.60e 5.00

e e 11.15 10.20e 12.30

e e e e

e e e e

Indraloris himalayensis Indraloris himalayensis Indraloris himalayensis Indraloris himalayensis

YPM 13802 GSI D237 PUA 688e69 Average

e e e e

e e e e

5.62 e e 5.62

2.42 e e 2.42

4.63 e e 4.63

e 6.65 5.34 6.00

e 2.91 2.44 2.68

e 5.81 4.29 5.05

e e 5.51 5.51

e e 2.38 2.38

e e 3.79 3.79

e e e e

e e 10.67 10.67

e e 12.14 12.14

Indraloris kamlialensis

YGSP 44443

e

e

(4.00)

e

3.34

e

e

e

e

e

e

e

e

e

Indraloris kamlialensis

YGSP 24338

4.68

2.71

e

e

e

e

e

e

e

e

e

e

e

e

Indraloris kamlialensis

Average

4.68

2.71

(4.00)

e

3.34

e

e

e

e

e

e

e

e

e

Indraloris sp. LARGE Indraloris sp. LARGE

YGSP 32152 Average

e e

e e

5.72 5.72

2.29 2.29

4.88 4.88

e e

e e

e e

e e

e e

e e

(14.0) e

e e

e e

Sinoadapis carnosus

PA 885 (left)

8.00

5.60

6.70

e

5.30

6.40

e

5.70

7.10

e

5.30

e

e

e

Sinoadapis carnosus

PA 885 (right)

e

e

6.20

e

5.50

6.30

e

5.70

e

e

e

16.30

e

e

Sinoadapis carnosus

Average

8.00

5.60

6.45

e

5.40

6.35

e

5.70

7.10

e

5.30

16.30

e

e

Sinoadapis shihuibaensis

PA 882

8.10

5.30

6.80

e

5.30

7.30

e

5.50

7.20

e

5.10

17.10

e

e

Sinoadapis shihuibaensis

Average

8.10

5.30

6.80

e

5.30

7.30

e

5.50

7.20

e

5.10

17.10

e

e

This study Chopra and Vasishat (1980) Chopra and Vasishat (1980) Vasishat (1985) This study

This study This study This study Flynn and Morgan (2005) Flynn and Morgan (2005) This study Pan and Wu (1986) Pan and Wu (1986) Pan and Wu (1986)

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

Sunetar sivaladapid

M3 MDTri

TF 6273

3.85

2.61

3.69

1.60

2.52

e

e

e

e

e

e

7.16

e

e

Siamoadapis maemohensis

TF 6233

4.01

2.68

3.52

e

2.53

e

e

e

e

e

e

6.58

e

e

Siamoadapis maemohensis

TF 6234

e

e

e

e

e

3.93

1.45

2.79

e

e

e

e

e

e

Siamoadapis maemohensis

Average

3.93

2.65

3.61

1.60

2.53

3.93

1.45

2.79

e

e

e

6.87

e

e

Guangxilemur singsilai

DBC 2170

e

e

5.07

e

3.51

e

e

e

e

e

e

e

e

e

Guangxilemur singsilai

DBC 2171

e

e

5.03

e

3.61

e

e

e

e

e

e

e

e

e

Guangxilemur singsilai

Average

e

e

5.05

e

3.56

e

e

e

e

e

e

e

e

e

Paukkaungia parva

NMMP 56

3.54

2.20

e

e

e

e

e

e

e

e

e

e

e

e

Paukkaungia parva

NMMP 55

e

e

3.59

e

2.45

e

e

e

e

e

e

e

e

e

Paukkaungia parva

NMMP 57

Paukkaungia parva

Average

Kyitchaungia takaii

Chaimanee et al. (2008); This study Chaimanee et al. (2008) Chaimanee et al. (2008); This study Marivaux et al. (2002) Marivaux et al. (2002) e Beard et al. (2007) Beard et al. (2007) Beard et al. (2007)

e

e

e

e

e

3.31

e

2.41

e

e

e

e

e

e

3.54

2.20

3.59

e

2.45

3.31

e

2.41

e

e

e

e

e

e

NMMP 28

e

e

e

e

e

4.39

e

(3.00)

e

e

e

e

e

e

Kyitchaungia takaii

Average

e

e

e

e

e

4.39

e

(3.00)

e

e

e

e

e

e

Wailekia orientale Wailekia orientale

TF 2632 Average

e e

e e

e e

e e

e e

4.60 4.60

1.60 1.60

3.60 3.60

4.70 4.70

3.20 3.20

(9.00) (9.00)

9.00 9.00

9.60 9.60

Yunnanadapis folivorus Yunnanadapis folivorus

IVPP V 22702 Average

4.30 4.30

2.76 2.76

4.56 4.56

e e

3.44 3.44

4.69 4.69

e e

4.09 4.09

5.08 5.08

e e

3.30 3.30

e e

e e

e e

Ni et al., 2016

Laomaki yunnanensis Laomaki yunnanensis Laomaki yunnanensis Laomaki yunnanensis Laomaki yunnanensis

IVPP V 22710 IVPP V 22711 IVPP V 22712 IVPP V 22713 Average

2.15 e e e 2.15

1.70 e e e 1.70

e 2.45 e e 2.45

e e e e e

e 2.00 e e 2.00

e e 2.65 e 2.65

e e e e e

e e 2.15 e 2.15

e e e 3.25 3.25

e e e e e

e e e 2.05 2.05

e e e e e

e e e e e

e e e e e

Ni Ni Ni Ni

Hoanghonius stehlini

Unnumbered, Uppsala

e

e

e

e

e

4.00

1.50

3.30

4.60

2.70

7.90

e

e

This study Tong et al. (1999)

Hoanghonius stehlini

IVPP 10220

3.40

2.30

3.90

e

2.80

4.00

e

3.00

4.60

Hoanghonius stehlini

Average

3.40

2.30

3.90

e

2.80

4.00

1.50

3.15

4.60

1.50 1.50

1.30 e 1.30

2.70

7.20

e

e

2.70

7.55

e

e

Rencunius zhoui

IVPP 5312

3.60

2.50

3.80

1.90

3.00

4.20

1.70

3.50

e

e

e

(7.90)

e

e

Rencunius zhoui

IVPP 5311

e

e

e

e

e

4.00

e

3.90

e

e

e

e

e

e

Rencunius zhoui

Specimen No. 3

e

e

e

e

e

e

e

e

4.50

e

2.90

e

e

e

Rencunius zhoui

Average

3.60

2.50

3.80

1.90

3.00

4.10

1.70

3.70

4.50

e

2.90

(7.90)

e

e

a b

Beard et al. (2007) This study

et et et et

al., al., al., al.,

2016 2016 2016 2016

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

Siamoadapis maemohensis

This study Gingerich et al. (1994) Gingerich et al. (1994); Woo and Chow (1957)

MD ¼ maximum mesiodistal length, MDTri ¼ maximum mesiodistal length of the trigonid, BL ¼ maximum buccolingual breadth, Mand. Ht. ¼ Mandibular Height. Numbers in parentheses represent estimates. Average and Range for S. nagrii MD, BL, and Mand Ht. measurements taken from Gingerich and Sahni (1984) with the addition of the PUA specimens listed here.

25

Taxon

Specimen number

26

Table 3 Comparative relative measurements of the mandible and lower dentition for VPL/RSP1 and other sivaladapid primates.a P4 MD/ M1 MD

M1 MD/ M1 BL

M1 MDTri/ M1 MD

M2 MD/ M2 BL

M2 MDTri/ M2 MD

M3 MD/ M3 BL

M3 MDTri/ Mand. Ht. at Mand. Ht. M1 Area/ M3 MD M1/M1 MD at M1/M2 MD M2 Area

1.18

1.38

0.29

1.27

0.36

1.40

0.36

1.81

1.79

M1 Area/ M3 Area

M2 Area/ M3 Area

0.89

0.90

1.00

VPL/RSP1

Sivaladapis palaeindicus

YGSP 25441

e

e

e

e

e

e

e

e

e

e

e

e

Sivaladapis palaeindicus Sivaladapis palaeindicus

GSI De224 YPM 19134

e e

e e

e e

1.28 e

0.40 e

1.42 1.38

0.31 0.30

e e

e e

e e

e e

0.76 e

Sivaladapis palaeindicus

BSPHG 1939 X5

e

e

e

1.39

e

e

e

e

e

e

e

e

Sivaladapis Sivaladapis Sivaladapis Sivaladapis Sivaladapis Sivaladapis Sivaladapis

YGSP 31881 YGSP 32150 YGSP 32154 YGSP 46458 YGSP 31720 YGSP Se394 AMNH 103370

e e e e e e e

e e 1.27 1.28 e e e

e e 0.37 0.41 e e e

e e e 1.29 e e e

e e e 0.38 e e e

1.41 1.57 e e e e 1.54

0.30 0.31 e e e e 0.30

e e e 1.82 e e e

e e e 1.74 e e e

e e e 0.92 e e e

e e e e e e e

e e e e e e e

palaeindicus palaeindicus palaeindicus palaeindicus palaeindicus palaeindicus palaeindicus

Sivaladapis palaeindicus

Unnumbered

1.06

1.36

e

e

e

e

e

e

e

e

e

e

Sivaladapis palaeindicus

Average

1.10b

1.30

0.39

1.32

0.39

1.46

0.30

1.82

1.74

0.92

0.83b

0.76

0.99

1.45

0.37

1.31

0.30

1.40

0.24

e

e

0.89

0.88

0.99

e e

e 1.23

e 0.35

e 1.22

e 0.32

e e

e e

e e

e e

e 0.82

e e

e e

0.79

0.80

1.01

e

e

1.08

Sivaladapis nagrii Sivaladapis nagrii

LUVP 14500/ 116860 LUVP 14503 GSI 18093 (TYPE)

Sivaladapis nagrii

PUA 826e69

1.02

1.26

e

1.40

e

1.50

e

e

e

Sivaladapis nagrii

PUA 343e69

e

e

e

1.24

e

1.38

e

e

e

Sivaladapis nagrii Sivaladapis nagrii Sivaladapis nagrii Sivaladapis nagrii

PUA 72e10 PUA 736e69 Averageb Rangeb

e e 0.24 0.24

e e 1.97b e

e e 1.77b e

Indraloris himalayensis Indraloris himalayensis Indraloris himalayensis Indraloris himalayensis

YPM 13802 GSI D237 PUA 688e69 Average

e e e e

1.21 e e 1.21

0.43 e e 0.43

e 1.14 1.24 1.19

e 0.44 0.46 0.45

e e 1.45 1.45

e e 0.43 0.43

e e e (1.90)

e e e (1.78)

Indraloris kamlialensis Indraloris kamlialensis Indraloris kamlialensis

YGSP 44443 YGSP 24338 Average

e e 1.17b

1.20 e 1.20

e e e

e e e

e e e

e e e

e e e

e e e

Indraloris sp. LARGE Indraloris sp. LARGE

YGSP 32152 Average

e e

1.17 1.17

0.40 0.40

e e

e e

e e

e e

Sinoadapis carnosus Sinoadapis carnosus Sinoadapis carnosus

PA 885 (left) PA 885 (right) Average

1.19 e 1.19

1.26 1.13 1.20

e e 0.33c

1.12 1.11 1.11

e e 0.3c

1.34 e 1.34

Sinoadapis shihuibaensis Sinoadapis shihuibaensis

PA 882 Average

1.19 1.19

1.28 1.28

e 0.36c

1.33 1.33

e 0.32c

Siamoadapis maemohensis

TF 6273

1.04

1.46

0.43

e

Siamoadapis maemohensis

TF 6233

1.14

1.39

e

Siamoadapis maemohensis

TF 6234

e

e

e

Sivaladapis nagrii

e 1.17 e 1.24 e e e 1.26 0.34 e e e 1.07b 1.22b 0.35 1.26b 0.31 1.46b 0.99e1.02 1.17e1.45 0.34e0.37 1.24e1.40 0.30e0.32 1.38e1.50

0.80 e e e e e 0.83b 0.80b 0.96b 0.79e0.89 0.80e0.88 0.99e1.08

This study Flynn and Morgan (2005) This study This study Flynn and Morgan (2005) This study This study This study This study This study This study This study Thomas and Verma (1979)

This study This study This study Chopra and Vasishat (1980) Chopra and Vasishat (1980) Vasishat (1985) This study

This study This study This study

e e e 0.86b

e e e 1.25b

e e 1.10 1.10

e e e

e e e

e e e

e e e

Flynn and Morgan (2005) Flynn and Morgan (2005)

(2.45) e

e e

e e

e e

e e

This study

e e 0.25c

e 2.63 2.63

e 2.59 2.59

0.97 0.95 0.96

0.94 e 0.94

0.97 e 0.97

Pan and Wu (1986) Pan and Wu (1986)

1.41 1.41

e 0.26c

2.51 2.51

2.34 2.34

0.90 0.90

0.98 0.98

1.09 1.09

Pan and Wu (1986)

e

e

e

1.94

e

e

e

e

e

e

e

e

1.87

e

e

e

e

1.41

0.37

e

e

e

e

e

e

e

b

b

Chaimanee et al. (2008); This study Chaimanee et al. (2008) Chaimanee et al. (2008); This study

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

Sunetar sivaladapid

Reference

Siamoadapis maemohensis Average

1.09

1.43

0.43

1.41

0.37

e

e

1.90

(1.75)

b

0.83b

e

e

DBC 2170 DBC 2171 Average

e e e

1.45 1.39 1.42

e e 0.34c

e e e

e e e

e e e

e e e

e e e

e e e

e e e

e e e

e e e

Marivaux et al. (2002) Marivaux et al. (2002) e

Paukkaungia parva Paukkaungia parva Paukkaungia parva Paukkaungia parva

NMMP 56 NMMP 55 NMMP 57 Average

e e e 0.99b

e 1.47 e 1.47

e e e 0.41c

e e 1.37 1.37

e e e 0.44c

e e e e

e e e e

e e e e

e e e e

e e e 1.10b

e e e e

e e e e

Beard et al. (2007) Beard et al. (2007) Beard et al. (2007)

Kyitchaungia takaii Kyitchaungia takaii

NMMP 28 Average

e e

e e

e e

1.46 1.46

e 0.41c

e e

e e

e e

e e

e e

e e

e e

Beard et al. (2007)

Wailekia orientale Wailekia orientale

TF 2632 Average

e e

e e

e e

1.28 1.28

0.35 0.35

1.47 1.47

0.32 0.32

e (2.10)

1.96 1.96

e e

e e

1.10 1.10

This study

Yunnanadapis folivorus Yunnanadapis folivorus

IVPP V 22702 Average

0.94 0.94

1.33 1.33

e 0.43c

1.15 1.15

e 0.36c

1.54 1.54

e 0.29c

e e

e e

0.82 0.82

0.94 0.94

1.14 1.14

Ni et al., 2016

Laomaki yunnanensis Laomaki yunnanensis Laomaki yunnanensis Laomaki yunnanensis Laomaki yunnanensis

IVPP V 22710 IVPP V 22711 IVPP V 22712 IVPP V 22713 Average

e e e e 0.88

e 1.23 e e 1.23

e e e e 0.31c

e e 1.23 e 1.23

e e e e 0.39c

e e e 1.59 1.59

e e e e 0.28c

e e e e e

e e e e e

e e e e e

e e e e e

e e e e e

e

e

e

1.21

0.38

1.70

0.28

e

1.98

e

e

1.06

This study

Hoanghonius stehlini Hoanghonius stehlini

Unnumbered, Uppsala IVPP 10220 Average

0.87 0.87

1.39 1.39

e 0.46c

1.33 1.27

e 0.38

1.70 1.70

e 0.28

1.85 1.85

1.80 1.89

0.91 0.91

0.88 0.88

0.97 1.01

Tong et al. (1994)

Rencunius zhoui Rencunius zhoui

IVPP 5312 IVPP 5311

0.95 e

1.27 e

0.50 e

1.20 1.03

0.40 e

e e

e e

2.08 e

1.88 e

0.78 e

e e

e e

Hoanghonius stehlini

Rencunius zhoui

Specimen No. 3

Rencunius zhoui

Average

e

e

e

e

e

1.55

e

e

e

e

e

e

0.95

1.27

0.50

1.11

0.40

1.55

0.29c

2.08

1.88

0.78

0.87b

1.16b

Ni Ni Ni Ni

et et et et

al., al., al., al.,

2016 2016 2016 2016

This study Gingerich et al. (1994) Gingerich et al. (1994); Woo and Chow (1957)

a MD ¼ maximum mesiodistal length, MDTri ¼ maximum mesiodistal length of the trigonid, BL ¼ maximum buccolingual breadth, Mand. Ht. ¼ Mandibular Height. Numbers in parentheses represent estimates. Sinoadapis estimates from Pan and Wu (1986), Pan (1988), and White (2006). Estimate for Rencunius zhoui taken from Woo and Chow (1957). All other estimates taken from the publication listed in the References column. b Where sample sizes were very low, value was taken by dividing the averages for each tooth rather than taking the average value from a number of individual specimens. c Trigonid compression estimate taken from published photographs.

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

Guangxilemur singsilai Guangxilemur singsilai Guangxilemur singsilai

27

28

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

University of Southern California's Molecular Imaging Center using a Nikon XT H 225 scanner with the following parameters: beam energy 170 kVp; beam current 38 mA; geometric magnification 11.437; voxel size 17.49 mm; and effective pixel resolution 57.1821 px/mm. Image stacks were saved as 16-bit unsigned images in DICOM format. Visualization of external and internal anatomy was facilitated with Amira v. 5.6 software (FEI, Inc.). A 3D rendering of the specimen is provided in the Supplementary Online Material (SOM) and the raw X-ray scan data is available on the Morphosource website (www.morphosource.org). In addition to micro-CT scans, the Sunetar specimen was ultrasonically cleaned, dried, and mounted on a stub for scanning electron microscopy. The partial mandible was first sputter coated with gold, then viewed and photographed using a JEOL SEM JSM 6490 housed at the Department of Geology, Panjab University.

2.2. Measurements and quantitative shape analyses Three measurements were taken on each lower molar of VPL/ RSP1 and the comparative fossil sample: maximum mesiodistal length (MD), maximum buccal-lingual breadth (BL), and the mesiodistal length of the trigonid (MDTri). MDTri was measured as the distance from the mesial margin of the tooth to the approximate midpoint of the protocristid connecting the protoconid and metaconid at the back wall of the trigonid. For each P4, MD and BL were also recorded. When possible, mandibular corpus depth (i.e., height) was measured on each specimen below each molar, M1-M3 (Table 2). For each upper molar (in the comparative fossil sample only), both MD and BL measurements were taken from the literature where available. Table 2 provides all raw linear measurements used in this study. To compare relative dental and mandibular shape among adapoids, a number of indices were created (Table 3). To estimate relative premolar length, P4 MD was divided by M1 MD. For each molar, MD/BL and MDTri/MD measure the overall shape of the tooth and the degree of trigonid compression for each lower molar, respectively. Additionally, the relative size of each molar was compared by using a ratio of tooth areas: M1/M2, M1/M3, and M2/ M3 for both upper and lower molars. Here, area was determined as the product of MD  BL in each tooth. Relative mandibular depth was assessed as the mandibular corpus depth at M1/M1 MD and/or the mandibular corpus depth at M1/M2 MD, depending on which measurement was available. In total, twelve relative measurements for all taxa preserving P4M3 were included in a series of statistical shape analyses (Table 3). Univariate comparisons were performed using t-tests and ANOVAs with post hoc comparisons (Tukey's HSD) to assess significant differences between taxa. In addition, the mean value for each measurement was calculated by species and subsets of these twelve measurements were then used in three multivariate Neighbor-Joining (NJ) cluster analyses to compare between as many taxa as possible. The first included 11 measurements representing all lower molar (M1-M3) indices and relative size ratios, mandibular depth at M1, and size (log M1 Area). The second analysis included seven measurements representing M1-M2 shape indices, M1/M2 relative size ratio, mandibular depth at M1, and size (log M1 Area). The third analysis included seven measurements representing M2-M3 shape indices, M2/M3 relative size ratio, mandibular depth at M2, and size (log M2 Area). The NJ analyses also allowed the assignment of an outgroup for dental and mandibular shape, in this case the primitive adapoid Donrussellia provincialis and a 9999 replication bootstrap procedure provided branch support. All statistical analyses were performed in PAST 3.10 (Hammer et al., 2001) and SPSS v22.0 (IBM Corp.).

2.3. Phylogenetic analysis In addition to qualitative and quantitative comparisons, we conducted a phylogenetic analysis of all known sivaladapid genera, largely building on the previous analyses by Qi and Beard (1998), with the addition of a few characters from Ni et al. (2016) and our own observations. All characters and character states are provided in Table 4 and the matrix is provided as a nexus file in the SOM. In total, 40 characters were included in the analysis, incorporating characters representing the upper dentition, lower dentition, and mandible. Where possible, characters were quantified and tested for any allometric influence by correlating shape indices against a body size proxy as described by Gilbert and colleagues (Gilbert and Rossie, 2007; Gilbert et al., 2009; Gilbert, 2013). In this case, estimated body mass was used as the proxy for body size; this allowed for some sivaladapid taxa known only from the upper dentition and others known only from the lower dentition to be placed on a common scale. For all those taxa preserving a lower M1, body mass was derived from the M1 area prosimian equations provided in Conroy (1987) or taken from Gilbert (2005; Table 5). Because the M2 area of Wailekia is almost exactly halfway between that of Rencunius and the Sunetar sivaladapid, the body mass of Wailekia was estimated by averaging the body mass of these two taxa. Kyitchaungia was estimated by taking the median of the estimated range given by Beard et al. (2007; i.e., 1040 g). Lushius was estimated at ~1450 g based on an M1 and M2 area slightly smaller than Guangxilemur singsilai, which is estimated on the basis of the M1 area prosimian equation at ~1526 g (Table 5). All allometrically influenced quantitative characters were coded using the general allometric coding method (see Gilbert et al., 2009; Gilbert, 2013) and all other quantitative characters were coded using gapweighted coding (Thiele, 1993), dividing the variation into three character states. Where it can be assumed that a population likely passed through an intermediate state to get to an extreme state on either side, ordering characters is a much more faithful representation of the evolutionary process (e.g., a population does not typically evolve from a small body mass to a large body mass without passing through an intermediate body mass through directional selection; see also Wiens, 2000). In this case, 37 characters were considered ordered, while three describing cristid obliqua orientation were left unordered (Table 4). All characters were left unscaled, meaning that each step in the analysis was weighted equally. The resulting matrix was analyzed using a furthest addition sequence Branch-and-Bound search to find the most parsimonious tree (MPT) in the software package PAUP* 4.10b (Swofford, 2003; see SOM for nexus file). To provide an estimate of clade support, a 1000 replication bootstrap procedure with replacement was performed, along with the calculation of decay indices/Bremer support values and the construction of the majority-rule consensus of all trees within 1% of the MPT (Strait et al., 1997; Gilbert, 2013). The primitive adapoids D. provincialis, Cantius (represented by Cantius torresi where available and C. ralstoni where characters for C. torresi could not be scored), and Marcgodinotius indicus were assigned and constrained as the outgroup, with the ingroup composed of all described sivaladapid taxa, Asiadapis (recently argued to be closely related or ancestral to sivaladapids by Godinot, 2015), the European cercamoniid genera Periconodon and Anchomomys1 and the

1
et al. (2016) have recently suggested that Anchomomys is not a Note that Marigo member of the Cercamoniidae, but is rather a close relative of azibiids, djebelemurids, and crown strepsirrhines on the basis of postcranial and dental data. We acknowledge this intriguing possibility but continue to follow the traditional classification of Anchomomys within Cercamoniidae for the purposes of this paper.

Table 4 Characters used in phylogenetic analysis.a Definition and character states

1 2 3 4 5

Lower molar hypoconulid shape; 0 ¼ absent/weak, 1 ¼ present and strong Lower molar hypoconulid position; 0 ¼ buccal, 1 ¼ central, 2 ¼ twinned with entoconid Lower molar lingual notch; 0 ¼ absent/weak, 1 ¼ present and strong Lower M1e2 paraconid development; 0 ¼ absent/indistinct, 1 ¼ polymorphic, 2 ¼ present/distinct Lower M1 cristid obliqua orientation; 0 ¼ distal to the protoconid, 1 ¼ near midline between protoconid and metaconid, 2 ¼ connected to the metaconid, 3 ¼ connected to protoconid near buccal border, 4 ¼ polymorphic between states 0 and 1 Lower M2 cristid obliqua orientation; 0 ¼ distal to the protoconid, 1 ¼ near midline between protoconid and metaconid, 2 ¼ connected to the metaconid, 3 ¼ connected to protoconid near buccal border Lower M3 cristid obliqua orientation; 0 ¼ distal to the protoconid, 1 ¼ near midline between protoconid and metaconid, 2 ¼ connected to the metaconid, 3 ¼ connected to protoconid near buccal border Entoconid-hypoconulid notch; 0 ¼ absent/weak, 1 ¼ present and strong Lower molar crown height; 0 ¼ moderate, 1 ¼ high-crowned P4 length relative to M1; 0 ¼ short, 1 ¼ long Upper and lower P4 molarization; 0 ¼ premolariform, 1 ¼ submolariform, 2 ¼ molariform P4 shape; 0 ¼ short/broad, 1 ¼ long/narrow Lower P4 paraconid development; 0 ¼ absent/weak, 1 ¼ present/strong Lower P4 talonid cusp development; 0 ¼ one distinct cusp present, 1 ¼ two distinct cusps present, 2 ¼ polymorphic between two and three cusps present, 3 ¼ three distinct cusps present M12 hypocone; 0 ¼ absent, 1 ¼ present and weak/small, 2 ¼ present and strong/large M12 pericone; 0 ¼ absent, 1 ¼ present and weak/small, 2 ¼ present and strong/large Lingual cingulum on upper molars; 0 ¼ weak/incomplete, 1 ¼ strong/continuous M12 shape; 0 ¼ square, 1 ¼ intermediate, 2 ¼ markedly transverse M3 shape; 0 ¼ square, 1 ¼ intermediate, 2 ¼ markedly transverse M12 molar conules; 0 ¼ distinct, 1 ¼ indistinct or absent M12 parastyle; 0 ¼ weak/absent, 1 ¼ strong M12 mesostyle; 0 ¼ weak/absent, 1 ¼ strong M12 external shearing crest shape; 0 ¼ mesiodistally straight, 1 ¼ moderately W-shaped, 2 ¼ W-shaped M1 area/M2 area; 0 ¼ M2 relatively large, 1 ¼ M1 relatively large M1 area/M3 area; 0 ¼ M3 relatively large, 1 ¼ intermediate, 2 ¼ M1 relatively large M2 area/M3 area; 0 ¼ M3 relatively large, 1 ¼ intermediate, 2 ¼ M2 relatively large Premolar number; 0 ¼ four, 1 ¼ polymorphic four or three, 2 ¼ three, 3 ¼ two P2 root number; 0 ¼ two, 1 ¼ one M1e2 shape; 0 ¼ short and broad, 1 ¼ intermediate, 2 ¼ long and narrow M1e2 trigonid compression; 0 ¼ mesiodistally compressed, 1 ¼ uncompressed M1e2 trigonid lingual shape; 0 ¼ lingually open, 1 ¼ polymorphic, 2 ¼ lingually closed Lower molar buccal cingulid strength; 0 ¼ absent/weak, 1 ¼ polymorphic/intermediate, 2 ¼ strong/continuous M3 shape; 0 ¼ short and broad, 1 ¼ long and narrow M3 trigonid shape; 0 ¼ compressed, 1 ¼ intermediate, 2 ¼ uncompressed M1 area/M2 area; 0 ¼ M2 relatively large, 1 ¼ intermediate, 2 ¼ M1 relatively large M1 area/M3 area; 0 ¼ M3 relatively large, 1 ¼ intermediate, 2 ¼ M1 relative M2 area/M3 area; 0 ¼ M3 relatively large, 1 ¼ intermediate, 2 ¼ M2 relatively large Mandibular height at M1; 0 ¼ shallow, 1 ¼ deep Symphyseal fusion; 0 ¼ mandibular symphysis unfused, 1 ¼ partially fused, 2 ¼ fused Estimated body mass; 0 ¼ small, 1 ¼ intermediate, 2 ¼ large

6 7 8 9 10b 11 12 13 14 15 16 17 18 19 20 21 22 23 24b 25 26 27 28 29 30b 31 32 33b 34 35 36 37 38b 39 40 a b c

Type QL, QL, QL, QL, QL,

Reference

O O O O U

Qi and Beard (1998) Qi and Beard (1998) c Qi and Beard (1998) This study; Ni et al., 2016 This study; Ni et al., 2016

QL, U

This study; Ni et al., 2016

c

QL, U

This study; Ni et al., 2016

c

QL, O QL, O QN, O QL, O QN, O QL, O QL, O

This study Qi and Beard (1998) This study Qi and Beard (1998) c This study This study; Ni et al., 2016 This study; Ni et al., 2016

QL, O QL, O QL, O QN, O QN, O QL, O QL, O QL, O QL, O QN, O QN, O QN, O QL, O QL, O QN, O QN, O QL, O QL, O QN, O QN, O QN, O QN, O QN, O QN, O QL, O QN, O

Qi and Beard (1998) c Qi and Beard (1998) c Qi and Beard (1998) Qi and Beard (1998) c This study Qi and Beard (1998) Qi and Beard (1998) Qi and Beard (1998) Qi and Beard (1998) c This study This study This study This study This study This study This study This study; Ni et al., 2016 This study This study This study This study This study This study This study This study This study

c

c c

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

Character

c

QL ¼ qualitative, QN ¼ quantitative, O ¼ ordered, U ¼ unordered. Indicates those characters influenced by allometry and coded using the general allometric coding method (Gilbert et al., 2009). All other quantitative characters coded using gap-weighted coding. Indicates that the character was modified from Qi and Beard (1998) or Ni et al. (2016), respectively.

29

30

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

Table 5 Body mass estimates for sivaladapid taxa.a Taxon VPL/RSP1 Sivaladapis palaeindicus Sivaladapis nagrii Indraloris himalayensis Indraloris kamlialensis Indraloris sp. Large Siamoadapis maemohensis Sinoadapis carnosus Sinoadapis shihuibaensis Hoanghonius stehlini Rencunius zhoui Wailekia orientale Guangxilemur tongi Guangxilemur singsilai Paukkaungia parva Kyitchaungia takaii Yunnanadapis folivorus Laomaki yunnanensis Lushius qinlinensis a

Subfamily

Geological age

Estimated average body mass (g)

Sivaladapinae Sivaladapinae Sivaladapinae Sivaladapinae Sivaladapinae Sivaladapinae Sivaladapinae Sivaladapinae Sivaladapinae Hoanghoniinae Hoanghoniinae Hoanghoniinae Incertae sedis Incertae sedis Incertae sedis Incertae sedis Incertae sedis Incertae sedis Incertae sedis

Middle Miocene Middle Miocene Late Miocene Late Miocene Middle Miocene Middle Miocene Middle Miocene Late Miocene Late Miocene Middle Eocene Middle Eocene Late Eocene Late Eocene Early Oligocene Middle Eocene Middle Eocene Early Oligocene Early Oligocene Late Eocene

1290 3426 2648 2779 948 3113 510 4449 4701 684 734 1012 4800 1526 483 1040 1228 188 1450

Body mass estimates derived from M1 area prosimian regression in Conroy (1987), except where noted in the text. See Materials and Methods for details.

caenopithecid genus Europolemur to provide consistency with the analysis of Qi and Beard (1998). General family and superfamily designations in this paper follow Fleagle (2013). In addition to parsimony, we also conducted a Bayesian analysis of the character matrix in MrBayes v3.2.6 (Ronquist et al., 2012). The Mk model for morphological data (Lewis, 2001) was employed with “coding ¼ variable” and “rates ¼ gamma”. The Dirichlet prior was fixed using “symdirihyperpr ¼ fixed (infinity)” so that all character states had equal frequency. Markov Chain Monte Carlo (MCMC) chains were run for 10 million generations, with two independent runs, three heated chains, and one cold chain each, temperature set at 0.065, sampling frequency set at every 1000 generations, and the burn-in set at 25 percent. Donrussellia was set as the Outgroup and Dorussellia, Cantius, and Marcgodinotius were constrained at the base of the tree. Using these settings, the average standard deviation of split frequencies was 0.004356, average PSRF for parameter values ¼ 1.000, and acceptance rates for swaps between chains ranged from 37 to 57% for both runs. The nexus file with the additional MrBayes block is also included in the SOM. Following the standard Bayesian analysis described above, we also ran a Bayesian “tip-dating” analysis in MrBayes v3.2.6. The tipdating analysis takes information regarding the ages of fossil taxa, rates of evolution, models of speciation, and models of extinction into account when constructing prior probability distributions (Sallam and Seiffert, 2016). Furthermore, as described by Sallam and Seiffert (2016), the tip-dating analysis provides age-estimates of fossil taxa based on their resulting phylogenetic position and the estimated rate of morphological evolution. In this case, the tipdating analysis provides a way to estimate the age of the sivaladapid from Sunetar, perhaps providing a more constrained age estimate in the absence of precise geochronological data from Sunetar and Ramnagar. Here, we used the assumptions described and published in Sallam and Seiffert (2016) with the following exceptions: the outgroup and topological constraints were the same as the first analysis, and thus the root node was not constrained to fall within a uniform prior; MCMC chains were run for 10 million generations; the temperature was set to 0.1; the “sampleprob” was set as close to 0 as possible (sampleprob ¼ 0.00000001), reflecting the fact that there are no extant taxa in the analysis; and the default consensus tree was constructed. Uniform priors were used for all taxa except Marcgodinotius, Asiadapis, and Indraloris kamlialensis; these three taxa were set as “fixed” due to the fact that they appear tightly constrained, with Marcgodinotius

and Asiadapis assigned an age of 54.5 Ma (Rose et al., 2009, 2014; Dunn et al., 2016) and I. kamlialensis assigned an age of 15.2 Ma (Flynn and Morgan, 2005). Because the program attempts to place the youngest taxon in the analysis close to “0” (assuming that at least one extant taxon is in the analysis), the fixed age of Marcgodinotius, constrained at the base of the tree, provides a way to recalibrate the analysis after the dates are generated by adding the difference between the Bayesian estimated age and the fixed age for this taxon to all other age estimates in the tree. Using these settings, the average standard deviation of split frequencies was 0.003412, average PSRF for parameter values ¼ 1.000, and acceptance rates for swaps between chains ranged from 34 to 49% over both runs. The nexus file for the tip-dating analysis is again provided in the SOM. 3. Results 3.1. Description of VPL/RSP1 VPL/RSP1 is a lower right mandibular fragment preserving the corpus below P4-M3, as well as the worn M1-M3 dentition (Figs. 3e5, SOM Figs. 1e4, see Tables 2 and 3 for available measurements). Unfortunately, only portions of the P4 mesial and distal roots are preserved (Figs. 3e6); the specimen is broken immediately anterior to the P4 mesial root. The P4 distal root has been displaced from its alveolar socket, most likely postmortem, but the inferior portion of the root remains fossilized, elevated superiorly and posteriorly from its original position (Fig. 6). The anteroinferior portion of the ascending ramus is preserved, illustrating a welldefined oblique line and a portion of the well-developed masseteric fossa. The depth of the mandibular corpus under the level of M1 and M2 is consistent around 8.5 mm (Table 3), with the corpus deepening posteriorly behind M2. Gonion is not preserved. VPL/RSP1 is recognized as a sivaladapid and distinguished from primitive adapoids by two distinctive features found in the lower molars: 1) well-developed lower molar hypoconulids, and 2) twinning of the lower molar entoconids and hypoconulids on the lingual sides of the teeth (Gingerich and Sahni, 1979, 1984; Qi and Beard, 1998; Gebo, 2002). The dentition of VPL/RSP1 is heavily worn, a feature that is also distinctive of many other Siwalik sivaladapid specimens, and likely relates to a diet including fibrous/ leafy material (e.g., Gingerich and Sahni, 1984; Flynn and Morgan, 2005; White, 2006; Patnaik et al., 2014). The high degree of wear

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

31

results in the exposure of dentine and shows that the enamel was very thin (Figs. 4 and 5); only remnants of the various cusps are preserved. However, the strong hypoconulid and twinned hypoconulideentoconid pair are clearly visible as defined dentine lakes in the posterolingual corners of M2-M3 (Fig. 4). Due to wear and postmortem damage, most cusps on M1 are not readily visible, although it can be safely assumed that the entoconid and hypoconulid were twinned on this tooth as well (Fig. 4). While only the roots of P4 are preserved, it is clear that P4 was two-rooted, and it can be deduced from the distance between the mesial and distal roots that P4 was a relatively long tooth, most likely longer than M1. Our estimates suggest that the length of P4 was ~5.2e5.6 mm (see also Fig. 6). Unfortunately, the exact degree of P4 molarization cannot be assessed, but the inferred length of P4 suggests that it was more likely than not to be a molariform or submolariform tooth based on comparisons with other sivaladapids and primitive adapoids. Sivaladapines, in particular, are noted to possess a molariform or submolariform P4, and as a consequence, the P4 is often as long or longer than the M1 behind it (Table 6). VPL/RSP1 appears to share this latter feature with sivaladapines, while more primitive adapoids and sivaladapids that lack molariform premolars are clearly separated from sivaladapines with a much shorter P4 (Table 6 and Fig. 7; ANOVA p < 0.01; Tukey's pairwise comparisons indicate Sivaladapinae is greater than Hoanghoniinae, Cercamoniidae þ Caenopithecidae, and Cantius, p < 0.05). Additional specimens preserving an undamaged P4 are needed to confirm this inference. M1 and M2 are heavily worn teeth that are relatively long and narrow, similar to most other sivaladapid taxa but different from Indraloris and Rencunius (Figs. 4 and 8, Table 3). The trigonids of M1M2 are relatively compressed mesiodistally, a similarity shared with Sivaladapis, Sinoadapis, G. singsilai, and possibly Wailekia and Laomaki (Fig. 9). The protocristid forms a strongly acute angle with the paracristid. The trigonid appears to have been lingually open, but the area is damaged in all three molars. Likewise, based on what is preserved, the paraconid appears to have been either absent or present only as a reduced cusp at the end of the paracristid, and this crest appears to have contacted the hypoconulid or postcristid of the preceding tooth, as described for Sivaladapis (Gingerich and Sahni, 1984; Flynn and Morgan, 2005). The cristid obliqua is oriented in a relatively buccal position in all of the molars, contacting the postvallid or protocristid wall just lingual to the protoconid, and creating a rounded angle with the hypocristid at the back of the tooth of ~90
(or slightly greater). Judging by the remaining dentine lakes, the hypoconulid and entoconid were prominent cusps on the lingual side of the molars. Notably, while there is evidence of a strong lingual notch between the posterior trigonid wall and the entoconid, there is no remaining evidence of a similarly strong notch or groove between the entoconid and hypoconulid as seen in Sivaladapis and many other sivaladapid taxa, although it is possible that this appearance is due to wear given the large dentine lake representing the hypoconulid. On the buccal side of the molars, the cingulid is relatively weak, only present as a restricted shelf between the protoconid and hypoconid, most similar to the condition seen in Indraloris, Wailekia, and Siamoadapis. Based on the prosimian M1 area regression in Conroy (1987), if VPL/RSP1 is representative of the population as a whole, the estimated body mass for the Sunetar sivaladapid is ~1290 g, notably

Figure 3. Photograph of VPL/RSP1 in A) buccal view. Three dimensional reconstruction of VPL/RSP1 created from high-resolution micro-CT scans: B) lingual, C) buccal, D) occlusal, E) inferior, F) posterior, G) anterior, and H) oblique views. Bright white areas on the buccal side of the body below M1 and M2, and on the occlusal surface of M1 are remnants of high density matrix not removed from the original specimen. Scale ¼ 1 cm. Photo Credit: Sheena Lad. See SOM for additional photos.

32

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

Figure 4. The dentition of VPL/RSP1 in: A-B) occlusal, C-D) buccal, and E-F) lingual views. Three-dimensional reconstructions from high-resolution micro-CT scans on the left (A, C, E). SEM images on the right (B, D, F). Note the distinctive features: 1) twinned entoconid-hypoconulid, 2) MD compressed trigonid, 3) weak buccal cingulid, 4) relatively deep lingual notch, 5) lack of distinct notch between entoconid and hypoconulid. Scale ¼ 1 mm. See SOM for additional photos.

Figure 5. Internal sagittal and coronal views of VPL/RSP1 from high-resolution micro-CT scans. (A) Illustration showing locations of two ‘sagittal’ slices (B, C) through all three lower molars and one coronal slice (D) at the level of M2. Note the thin enamel on the dentition of VPL/RSP1.

smaller than Sivaladapis, Sinoadapis, Guangxilemur, and Indraloris himalayensis, but almost double the estimated body mass of Rencunius, Hoanghonius, Paukkaungia, and Siamoadapis. Among described sivaladapid taxa, it is most similar in estimated body size to Yunnanadapis, Wailekia, Kyitchaungia, and I. kamlialensis (Table 5). We compared available dental and mandibular measurements of VPL/RSP1 with those taken from casts and original sivaladapid specimens in museum collections, as well as the literature, to examine phenetic similarities and differences in overall lower

molar shape. While most sivaladapid genera are only represented by a couple of described specimens at most, among the betterknown Siwalik genera our results suggest that Sivaladapis can be reliably distinguished from Indraloris by its significantly narrower molars (t-test of pooled M1-M2 MD/BL, p < 0.05) and significantly more compressed molar trigonids (t-test of pooled M1-M2 MDTri/ MD, p < 0.05; Table 3, Figs. 8e9). For M1-M3, specimen VPL/RSP1 either falls close to or within the range of Sivaladapis for these two molar features and outside of the Indraloris range.

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

33

Figure 6. Internal sagittal and coronal views of the P4 roots of VPL/RSP1 from high-resolution micro-CT scans. (A) Illustration showing locations of the ‘sagittal’ slice (B) through both mesial and distal roots of P4, and one coronal slice (C) at the level of the mesial root of P4 (shown in anterior view). A zoomed in region of the P4 roots in ‘sagittal’ view (D) illustrates a line drawing estimate of P4 to scale with a MD length ~5.2 mm. Note the broken distal root is preserved outside of its socket and projects posteriorly rather than superiorly. Scale ¼ 1 cm. See SOM for additional photos.

The three iterations of NJ cluster analyses provide a broader phenetic assessment, including as many combinations of sivaladapid taxa as possible. They indicate that in terms of overall shape, VPL/RSP1 is most similar to sivaladapine taxa such as

Table 6 P4 MD/M1 MD ratio among sivaladapid and early adapoid taxa.a Taxon

P4 MD/M1 MD Range

VPL/RSP1

1.10e1.18

This study

Sivaladapinae Hoanghoniinae Donrussellia provincialis Average Marcgodinotius indicus Average Asiadapis cambayensis (n ¼ 2) Cantius torresi (n ¼ 4)

0.99e1.19 0.87e0.95 0.92 0.96 0.91e0.97 0.83e0.91

Europolemur klatti Average

0.96

Anchomomys quercyi (n ¼ 1) Periconodon huerzeleri (n ¼ 1) Paukkaungia parva Average Yunnanadapis folivorus Laomaki yunnanensis

0.94 0.90 0.99 0.94 0.88

see Table 2 see Table 2 Godinot, 1981 Rose et al., 2009 Rose et al., 2009 Gingerich, 1986 Gingerich, 1995 Godinot, 1988a; Gilbert, 2005 Godinot, 1988b Godinot, 1988a see Table 2 see Table 2 see Table 2

a

Reference

MD ¼ maximum mesiodistal length, Average ¼ value calculated by taking the average P4 MD value and average M1 MD value from the reference provided. For VPL/RSP1, a range is provided based on an estimated P4 MD between 5.2 and 5.6 mm.

Sivaladapis, Sinoadapis, and Indraloris, yet also distinct and perhaps slightly more primitive than these taxa (Fig. 10). The VPL/ RSP1 þ sivaladapine grouping was recovered in all three NJ analyses, suggesting that the Sunetar specimen is closely related to these taxa and derived in comparison to the more primitive hoanghoniines. Given the consistent morphological evidence that VPL/RSP1 represents a primitive and distinct sivaladapine, we formally place the specimen in a new genus and species.2

3.2. Systematic paleontology Order Primates Linnaeus, 1758
Geoffroy, 1812 Semiorder Strepsirrhini E. Infraorder Adapiformes Hoffstetter, 1977 Superfamily Adapoidea Trouessart, 1879 Family Sivaladapidae Thomas and Verma, 1979 Subfamily Sivaladapinae Thomas and Verma, 1979 Ramadapis sahnii gen. et sp. nov.

2 This published work and the nomenclatural acts it contains have been registered in ZooBank, the official online register of the ICZN. ZooBank Life Science Identifiers (LSIDs) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix ‘‘http:// zoobank.org/’’. The LSID for this publication is: urn:lsid:zoobank.org:act: 131269FC-6A12-469E-9B4F-4874133E969F.

34

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

Figure 7. P4 length relative to M1 length (P4 MD/M1 MD). Solid black dots represent individual fossil specimen values. Vertical black lines represent the known observed range for each taxon or taxonomic group. Data for VPL/RSP1 is based on the two estimates described in the text. Note that VPL/RSP1 shares a relatively long P4 exclusively with other sivaladapine taxa. See also Table 6.

3.3. Etymology Genus is in reference to the town of Ramnagar in combination with the genus Adapis, the type genus of the superfamily Adapoidea. The species name honors Professor Ashok Sahni, in recognition

Figure 8. Pooled M1 and M2 shape ratios (i.e., M1 MD/BL and M2 MD/BL, respectively). Solid black dots represent individual fossil specimen values. Vertical black lines represent the known observed range for each taxon. VPL/RSP1 has relatively longer than wide M1 and M2. See also Tables 2 and 3.

Figure 9. Pooled M1 and M2 trigonid compression ratios (i.e., M1 MDTri/M1 MD and M2 MDTri/M2 MD, respectively). Solid black dots represent individual fossil specimen values. Vertical black lines represent the known observed range for each taxon. VPL/ RSP1 has relatively compressed M1 and M2 trigonids. See also Tables 2 and 3.

of his many contributions to Indian and Siwalik paleontology, including the first identification of Sivaladapis and Indraloris as adapoids (Gingerich and Sahni, 1979). 3.4. Generic diagnosis Ramadapis is a medium-sized sivaladapid possessing relatively long and narrow lower molars with mesiodistally compressed trigonids and lacking evidence for a distinct notch between the twinned entoconid-hypoconulid. Molar crowns are moderate in height and enamel is thin. Buccal cingulids on the lower molars are weak and the cristid obliqua is oriented in a relatively buccal position, contacting the postvallid/protocristid wall just lingual to the protoconid. P4 is relatively long compared to the M1 behind it. Ramadapis is distinguished from all hoanghoniines and Laomaki by its larger size, and also distinguished from these taxa as well as Yunnanadapis by a relatively long P4 compared to the M1. It is further separated from hoanghoniines (and Kyitchaungia) by the lack of a closed trigonid and from Hoanghonius, Rencunius, Paukkaungia, and Kyitchaungia by a relatively deep lingual notch on the lower molars. With the exception of Indraloris, Paukkaungia, and Kyitchaungia, Ramadapis differs from all other described sivaladapids in its apparent lack of a well-defined notch between the entoconid and hypoconulid. The weak buccal cingulids found on Ramadapis lower molars differentiate it from most other sivaladapid taxa except Indraloris, Siamoadapis, Wailekia, and Kyitchaungia. Ramadapis most clearly differs from Indraloris in its long and narrow molar shape with mesiodistally compressed trigonids. Compared to Sivaladapis and Sinoadapis, Ramadapis is ~50e75% smaller and exhibits molar crowns with much weaker buccal cingulids. In addition to the lack of an entoconid-hypoconulid notch, Ramadapis differs from Siamoadapis in being almost twice as large and possessing a relatively shallow mandible, relatively compressed M1-M2 trigonids, relatively broader M1-M2, and a slightly smaller M2 relative to M1.

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

35

3.5. Specific diagnosis As for genus. 3.6. Holotype VPL/RSP1, a partial right mandible preserving the corpus under P4-M3, the roots of P4, and the worn crowns of M1-M3. 3.7. Hypodigm The holotype is the only known specimen. 3.8. Horizon Chinji-age Lower Siwalik deposits, ~14e11 Ma. 3.9. Localities/sites Sunetar site complex, just outside of the town of Ramnagar, Jammu and Kashmir State, India. 3.10. Phylogenetic analysis Results of a phylogenetic analysis including 40 dental and mandibular characters are not well-resolved. Allometry does not have a particularly strong influence on the overall analysis, with only five characters out of 40 noted as allometrically influenced. Parsimony analyses recover 62 MPTs, the majority of which support the position of R. sahnii as a distinct new sivaladapid taxon near the base of the sivaladapine radiation (Fig. 11; see .tre file in the SOM for individual MPTs). The parsimony strict consensus and the Bayesian consensus are similar in that they find strong support only for a Sivaladapis þ Sinoadapis clade, with the Bayesian consensus and parsimony 1% tree also finding support for an Indraloris clade (Figs. 11aeb,12). The parsimony analysis is more resolved, with all consensus trees finding support for a basal group of sivaladapids consisting of Hoanghonius, Laomaki, Paukkaungia, and Kyitchaungia, as well as a more derived clade consisting of Wailekia, Guangxilemur tongi, Ramadapis, Siamoadapis, Indraloris, Sivaladapis, and Sinoadapis (Fig. 11). The majority-rule consensus supports a further derived sivaladapine clade consisting of Siamoadapis, Ramadapis, Indraloris, Sivaladapis, and Sinoadapis, with Ramadapis in an unresolved position relative to the other Siwalik genera þ Sinoadapis (Fig. 11c). Bootstrap support values and Bremer support values are both very low: no clades outside of Sivaladapis þ Sinoadapis were supported more than 50% of the time and the decay index at every node was one (Fig. 11a). This is perhaps to be expected for any dataset with a small number of characters and a number of taxa with large amounts of missing data. Nevertheless, the majority-rule consensus of 53,418 trees within 1% of the MPTs provides some additional support for a close relationship between Sinoadapis and Sivaladapis and also supports a monophyletic Indraloris (Fig. 11b) within a broader Sivaladapine þ Wailekia þ G. tongi clade. As suggested in the majority-rule consensus tree, the lack of resolution among sivaladapines þ Wailekia þ G. tongi is largely due to the varying positions of the latter two taxa (Fig. 11c), and G. tongi is particularly unstable. Both taxa are represented by only two published teeth and present a mix of characters shared with earlier hoanghoniines and later sivaladapines (Ducrocq et al., 1995; Qi and Beard, 1998; Marivaux et al., 2002). While the Oligocene G. singsilai Figure 10. Results of NJ analyses of molar and mandibular shape. (A) Analysis with M1 and M2, (B) analysis with M2 and M3, and (C) analysis with M1-M3. Numbers represent bootstrap values for each corresponding branch.

36

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

Figure 11. Summary of 62 MPTs recovered from an analysis of 40 dental and mandibular characters (see Table 4). Tree length ¼ 182, CI ¼ 0.3571, RI ¼ 0.5923, RC ¼ 0.2115, HI ¼ 0.6429. a) Strict consensus tree of the 62 MPTs; bootstrap support values above 50% are provided above a given branch, decay indices are provided in parentheses below each branch on the tree. b) Majority-rule consensus tree of all trees within 1% of the MPT. Percentage of trees within 1% of the MPT found supporting a given clade is provided below or above each branch on the tree. c) Majority-rule consensus tree of the 62 MPTs. Percentage of trees found supporting any given branch is provided below or above each branch.

is placed just outside of the base of the sivaladapine radiation in all MPTs, the position of the Eocene G. tongi varies greatly from a branching position just after G. singsilai, to a branching position just after Wailekia, to a position as the sister taxon of Wailekia, and finally to multiple positions within Sivaladapines (Fig. 11). Many of these arrangements question the integrity of the genus Guangxilemur and some of the MPTs suggest that G. tongi and Wailekia orientale may be congeneric, although they do not share any overlapping dental

material (a close relationship between Guangxilemur and Wailekia is also discussed by Marivaux et al., 2002). The derived shearing crest morphology seen in G. tongi links this species with later Sivaladapis and Sinoadapis within Sivaladapines in at least two of the MPTs; it is linked with I. himalayensis and Indraloris sp. Large in a number of the MPTs based on the fact that they share a similar estimated body mass. The Bayesian analysis finds a monophyletic Sivaladapidae, but the relationships within the group are unresolved outside of a

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

37

Figure 12. Results of Bayesian analyses of 40 dental and mandibular characters (see Table 4). Top: Consensus tree resulting from standard Bayesian analysis; posterior probabilities supporting a given clade are provided below each branch on the tree. Bottom: Consensus tree resulting from Bayesian tip-dating analysis; posterior probabilities supporting a given clade are provided below or below each branch on the tree. X-axis in millions of years; colors correspond to relative geological age from red (older) to brown (younger). See Table 7 for mean age estimates and 95% credibility age intervals estimated for each taxon. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

38

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

monophyletic Indraloris and a Sivaladapis þ Sinoadapis clade (Fig. 12). One notable difference between the Bayesian and parsimony analyses concerns the placement of Lushius; while the parsimony analysis places Lushius near the base of the tree, outside of the European adapoids and Sivaladapidae, the Bayesian analysis includes Lushius within a monophyletic group including the other recognized members of Sivaladapidae. Both the parsimony and Bayesian analyses reconstruct Asiadapis at the base of the Ingroup, outside of Sivaladapidae þ European adapoids (Figs. 11 and 12); therefore, the evidence presented here is in contradiction to the recent hypothesis by Godinot (2015) suggesting Asiadapis is closely related to later sivaladapids. The Bayesian tip-dating analysis, taking rates of evolution and ages of fossil specimens into account, presents a tree comparable to the parsimony analyses in that it is more resolved and recovers a monophyletic Sivaladapinae including Siamoadapis, Ramadapis, Indraloris, and Sivaladapis þ Sinoadapis (Fig. 12). However, the influence of the age prior on the relationships recovered is readily apparent as the branching pattern mostly recovers Eocene adapoid taxa at the unresolved base of the tree (except the Oligocene Laomaki), followed by the Oligocene taxa Yunnanadapis and G. singsilai, and finally the Miocene sivaladapine clade. The reconstructed 95% probability age range for R. sahnii and Sunetar based on the tip-dating analysis is 11.2e14.1 Ma, equivalent to the age of the Chinji Formation and in line with our previous assessment based on limited faunal comparisons (Table 7; see also Gilbert et al., 2014). Compared with previous analyses, the parsimony strict consensus and majority-rule consensus trees are roughly similar to that suggested by Qi and Beard (1998) in that they hypothesize a paraphyletic Hoanghoniinae, with Hoanghonius, Rencunius, and Wailekia representing successive branching events prior to the branching of sivaladapines (Fig. 11). They are also similar to trees recently reported by Ni et al. (2016) in that they place Laomaki, Hoanghonius, and Paukkaungia close to the base of the sivaladapid radiation, although in the current analysis Laomaki is regarded as the sister taxon to Hoanghonius, whereas Ni et al. (2016) find it as

Table 7 Estimated mean ages and 95% confidence intervals for taxa included in the Bayesian tipedating analysis, in millions of years (Ma). Taxon Donrussellia provincialis Canitus torresi Marcgodinotius indicus Asiadapis cambayensis Europolemur klatti Anchomomys sp. Periconodon huerzeleri Rencunius zhoui Hoanghonius stehlini Wailekia orientale Guangxilemur tongi Guangxilemur singsilai Sivaladapis palaeindicus Sivaladapis nagrii Sinoadapis carnosus Sinoadapis shihuibaensis Ramadapis sahnii Indraloris himalayensis Indraloris kamlialensis Indraloris sp. Large Lushius qinlinensis Siamoadapis maemohensis Paukkaungia parva Kyitchaungia takaii Yunnanadapis folivorus Laomaki yunnanensis

Mean age estimate (Ma)

Lower 95% HPD

Upper 95% HPD

54.979704 55.641059 54.500000 54.500000 45.974183 42.298027 44.006471 39.840867 36.528863 35.966140 35.886782 30.210912 12.562954 8.919854 6.514779 6.604473 12.757317 8.937215 15.200000 15.953780 35.862277 13.199762 37.131054 36.854625 33.240692 33.270453

53.240100 55.275334 54.190238 54.190238 43.610789 38.932572 43.399581 37.791402 35.138705 34.960760 34.931586 25.930692 11.235904 8.462516 6.443154 6.443154 11.201117 8.477468 14.890243 15.116681 33.928222 12.856377 33.937605 33.896475 32.538169 32.529969

56.022596 55.982177 54.743065 54.743065 48.458564 44.668301 44.631954 42.221574 37.888017 37.053989 37.011340 33.164911 14.519850 9.392712 6.828336 6.968082 14.105469 9.396679 15.443069 16.852623 37.880482 13.499091 40.794686 40.331430 34.012465 34.006433

intermediate between Hoanghonius and a basal Rencunius. The Middle Eocene Kyitchaungia and Paukkaungia are consistently recovered as sister taxa closely related to Hoanghonius and Laomaki (Fig. 11). The recently described Oligocene taxon Yunnanadapis folivorus is most often placed in an intermediate position between Rencunius and later sivaladapines, along with the genera Wailekia and Guangxilemur (Fig. 11), again similar to its position in the Ni et al. (2016) analysis. Taken in sum, the phylogenetic hypothesis preferred here is summarized in the majority-rule consensus tree of the parsimony analysis (Fig. 11). Many of these trees suggest, in some combination, G. singsilai, Yunnanadapis, G. tongi, and Wailekia as successive branches just outside the base of the sivaladapine radiation (Fig. 11). After Yunnanadapis, Guangxilemur, and Wailekia, this hypothesis reconstructs Siamoadapis as the next taxon to branch off (as suggested by Chaimanee et al., 2008), followed by Ramadapis, Indraloris, and Sinoadapis þ Sivaladapis in some combination. 4. Discussion Despite almost 100 years of paleontological collection and research in the Ramnagar area, our investigations demonstrate that the region is still productive and holds the potential to increase our knowledge of primate and non-primate mammalian evolution. Ramadapis sahnii gen. et sp. nov. adds to the diversity of Siwalik primates known from this time period and adds to our understanding of the sivaladapid radiation, particularly the Miocene sivaladapines. R. sahnii appears most similar in overall dental shape and features to the sivaladapine taxa Sivaladapis, Indraloris, and Sinoadapis (Figs. 3e7), particularly Sivaladapis in its thin enamel, relatively long and narrow M1-M2, and mesiodistally compressed trigonids. It most obviously differs from Sivaladapis in its smaller size, weaker buccal cingulids, relatively smaller M3, and apparent lack of a well-marked sulcus between the entoconid and hypoconulid (although it is possible that future specimens with less wear will demonstrate that a deep entoconid-hypoconulid notch did typically exist). In sum, R. sahnii presents a primitive and generalized sivaladapine mandibular and dental morphology that, along with Siamoadapis, could easily approximate the ancestral condition for the classic Miocene genera, Indraloris, Sivaladapis, and Sinoadapis. Since the initial recognition of Indraloris and Sivaladapis as adapoids in 1979 (Gingerich and Sahni, 1979; Thomas and Verma, 1979), numerous taxa have subsequently been discovered and/or recognized as closely related forms and placed in the broader family-level group Sivaladapidae (e.g., Wu and Pan, 1985; Pan and Wu, 1986; Gingerich et al., 1994; Ducrocq et al., 1995; Qi and Beard, 1998; Marivaux et al., 2002; Flynn and Morgan, 2005; Beard et al., 2007; Chaimanee et al., 2008; Ni et al., 2016). Qi and Beard (1998) made an early attempt to clarify relationships among sivaladapids, and more recently Ni et al. (2016) attempted to place a few sivaladapid taxa within the larger primate radiation. While our analysis does not examine as many characters as some more recent analyses on adapoids and early primates as a whole (e.g., Seiffert et al., 2005, 2009, 2015; Rose et al., 2009; Marivaux
et al., 2016; Ni et al., 2016), it is notable in et al., 2013; Marigo that it includes nearly all taxa that have been previously argued to be sivaladapids and also includes some of the most basal taxa of other adapoid groups for comparison. Thus, the phylogenetic analysis performed here represents the most comprehensive attempt to clarify relationships among known sivaladapid taxa to date. Our results (Figs. 11e12) suggest that R. sahnii is indeed most likely a primitive sivaladapine, placed near the base of a clade containing Siamoadapis, Indraloris, and Sivaladapis þ Sinoadapis in some combination. The two taxa represented by the most material, Sivaladapis and Sinoadapis, are strongly supported as sister taxa.

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

Support for the broader sivaladapine clade is low, however, reflecting the fact that sivaladapids remain poorly known and relatively poorly sampled in terms of overall anatomy. Nevertheless, based on our analyses, some new hypotheses regarding sivaladapid systematics and evolution can be discussed. For example, the poorly known Eocene taxa Lushius, Paukkaungia, Kyitchaungia, and G. tongi have been previously described as sivaladapids, but generally classified as incertae sedis within the family. Our phylogenetic analysis suggests the possibility that Lushius may not belong to the family Sivaladapidae; it is placed outside of all other sivaladapids and European adapoids in the parsimony analysis, yet grouped within a monophyletic Sivaladapidae in the Bayesian analysis. Kyitchaungia and Paukkaungia are reconstructed as primitive sivaladapid taxa in our analysis, perhaps closely related to Hoanghonius and Laomaki. Along with Yunnanadapis and Wailekia, G. tongi and G. singsilai, which have not been previously placed within a sivaladapid subfamily, are most often positioned just outside of Sivaladapinae, between Rencunius and Siamoadapis. We see several possibilities for the classification of Guangxilemur, Yunnanadapis, and Wailekia moving forward. First, if G. singsilai and G. tongi are distantly related, as suggested in many of the MPTs (and hinted at by Marivaux et al., 2002), then G. singsilai may need a new genus. The larger question, however, concerns the attribution of these taxa to subfamily, along with Yunnanadapis and Wailekia. Including both Guangxilemur species, Yunnanadapis, and Wailekia within Sivaladapinae would necessitate a morphological expansion of the subfamily, making the only obvious difference between Sivaladapinae and Hoanghoniinae reliant on the presence of a long, molariform or submolariform upper and lower P4 in the former and absence in the latter. Exclusion of these three taxa from Sivaladapinae results in the sivaladapines being diagnosable on the basis of reduced-to-absent pericones and hypocones on M12, as well as the presence of relatively long and/or molarized premolars (particularly upper and lower P4). If Wailekia, Guangxilemur, and Yunnanadapis are indeed outside of a clade containing Siamoadapis, Ramadapis, Indraloris, Sinoadapis, and Sivaladapis (Sivaladapinae), as suggested by the majority-rule consensus tree and our preferred MPTs, it is perhaps best to classify them separately in recognition of the obvious differences compared to sivaladapine taxa, particularly in the presence of a distinct pericone and hypocone on the upper molars. However, these three genera also appear derived relative to the Hoanghoniinae, most obviously in the reduced to absent conules present on the upper molars, the increased molarization of the upper and lower P4, and the reduction of the paraconid on the lower molars (still retained in many hoanghoniines). Thus, it might also be best to recognize Wailekia, Guangxilemur, and Yunnanadapis within their own gradistic subfamily, in which case Wailekia has priority (i.e., subfamily Wailekiinae including Guangxilemur, Yunnanadapis, and Wailekia; see also Marivaux et al., 2002; Fig. 11). Given the fragmentary nature of both Guangxilemur taxa and Wailekia, additional fossils will be necessary to decide between these competing alternatives. Similar to the previous analyses of Qi and Beard (1998) and Ni et al. (2016), we find that the previously recognized subfamily Hoanghoniinae is likely a paraphyletic assemblage of primitive Eocene sivaladapid taxa. Our analysis also suggests that Laomaki, Paukkaungia, and Kyitchaungia should be provisionally included in this concept of Hoanghoniinae as an assemblage of primitive sivaladapids. Within Sivaladapinae, Siamoadapis is most likely basal to a more derived group containing Indraloris, Sivaladapis, Ramadapis, and Sinoadapis (Fig. 11). Sivaladapis and Sinoadapis appear to be very closely related and possibly represent sister taxa, as suggested by previous studies (Pan and Wu, 1986; Pan, 1988; Qi and Beard, 1998; Ni et al., 2016).

39

Although we believe our analyses represent an advance over previous attempts to examine sivaladapid phylogeny, there is weak support for many of the recovered clades and ample room for improvement in the future. For consistency with previous analyses, we intentionally restricted ourselves to a relatively small number of taxa and characters. Future analyses may benefit from a much larger set of characters and taxa than examined here, so that sivaladapids can be considered within a broader context of adapoid and primate evolution. One suggestion might be to include all sivaladapid taxa into the most current matrix of Seiffert and colleagues
et al., 2016) or the recent analysis (e.g., Seiffert et al., 2015; Marigo by Ni et al. (2016). An open question remains as to the relationship of Sivaladapidae relative to other adapoid families and subfamilies; while our parsimony analysis largely considers them as closely related to the European cercamoniids (in particular the genus Periconodon, which shares prominent upper molar pericones, hypocones, and an unmolarized P4/P4 with Hoanghonius and Rencunius), this is in part due to the fact that basal notharctids and Marcgodinotius are assigned and constrained as part of the outgroup. Asiadapis was assigned to the Ingroup, however, and in contrast to the recent suggestion by Godinot (2015) that this taxon may be closely related to sivaladapids, our analyses uniformly placed Asiadapis at the very base of the tree. Limited postcrania attributed to Hoanghonius include a first metatarsal, similar in morphology to Cantius (Gebo et al., 1999; Patel et al., 2012), and distal phalanges similar to those described for notharctids (Maiolino et al., 2012; Gilbert and Maiolino, 2015) and caenopithecids (Koenigswald, 1979; Koenigswald et al., 2012) in possessing a grooming claw/grooming nail on the second pedal digit (Gebo et al., 2015), a characteristic likely to have been shared widely among adapoids. A proximal femur and fragmentary tarsal bones attributable to Kyitchaungia look generally similar to early notharctids such as Cantius (Beard et al., 2007; Marivaux et al., 2008). A humerus alternatively attributed to an amphipithecid (e.g., Ciochon et al., 2001; Ciochon and Gunnell, 2002) or an as yet unnamed (and dentally nonexistent) sivaladapid (Beard et al., 2007) from the Eocene Pondaung Formation in Myanmar is overall very similar to those known from notharctids as well, perhaps suggesting a close relationship between sivaladapids and notharctids. Given the fragmentary, though growing, nature of the sivaladapid fossil record, the origins of the Sivaladapidae are difficult to determine at this time and the continued search for additional, more complete fossils will be central to resolving the debate. In the meantime, to help estimate higher-level relationships, a slightly broader analysis considering additional characters and basal notharctids, asiadapids, cercamoniids, caenopithecids, and adapids as part of the ingroup will be necessary. Based on faunal comparisons (Brown et al., 1924; Pilgrim, 1927; Colbert, 1935; Gregory et al., 1938; Vasishat et al., 1978; Gaur and Chopra, 1983; Nanda and Sehgal, 1993; Basu, 2004; Parmar and Prasad, 2006; Sehgal and Patnaik, 2012; Patnaik, 2013; Gilbert et al., 2014; Parmar et al., 2016), the Ramnagar region in general appears to be Chinji-age equivalent, ~14e11 Ma, which places the first appearance date (FAD) of R. sahnii slightly later than the FAD of Indraloris and Sivaladapis in the Kamlial Formation (~18e14 Ma). Unfortunately, our Bayesian tip-dating analysis is no more precise, but it is consistent and also suggests an age for R. sahnii between 14.1 and 11.2 Ma. This suggests that earlier and morphologically primitive sivaladapine taxa similar to R. sahnii remain to be discovered, as perhaps hinted at by G. singsilai in the Oligocene of Pakistan (Marivaux et al., 2002) and the newly discovered Yunnanadapis and Laomaki in the Oligocene of China (Ni et al., 2016). Future investigations at Ramnagar and elsewhere will be crucial in

40

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41

documenting new taxa and additional specimens to flesh out a more complete picture of sivaladapid evolution in South Asia.

Acknowledgments We thank Judy Galkin (AMNH), Chris Norris and Daniel Brinkman (YPM), and Larry Flynn (Harvard) for access to original specimens and casts in their care. Susan Bell and the AMNH archives provided key correspondence between Barnum Brown and W.D. Matthew documenting the early AMNH exploration of the Ramnagar area. Yaowalak Chaimanee kindly provided casts of Siamoadapis for study. We also thank Larry Flynn and John Barry for helpful discussions and advice. Sheena Lad provided the photographs of VPL/RSP1 in Figure 3 and the SOM. Erik Seiffert, Laurent Marivaux, an anonymous reviewer, and the editors at JHE provided helpful comments that greatly improved this manuscript. Kelsey Pugh provided helpful advice and software instruction for MrBayes. Finally, we thank Bino Varghese and the University of Southern California's Molecular Imaging Center for access and assistance with micro-CT scanning. This study was generously supported by the Wenner-Gren Foundation, the PSC-CUNY faculty research award program, Hunter College, the AAPA Professional Development Program, the University of Southern California, and the Institute of Human Origins (ASU). RP is supported by Ministry of Earth Science project (MoES/P.O. (Geoscience)/46/2015) and DST, PURSE.

Supplementary Online Material Supplementary online material related to this article can be found at http://dx.doi.org/10.1016/j.jhevol.2016.10.001.

References Basu, P.K., 2004. Siwalik mammals of the Jammu Sub-Himalaya, India: an appraisal of their diversity and habitats. Quatern. Int. 117, 105e118. Beard, K.C., Marivaux, L., Tun, S.T., Soe, A.N., Chaimanee, Y., Htoon, W., Marandat, B., Aung, H.H., Jaeger, J.eJ., 2007. New sivaladapid primates from the Eocene Pondaung Formation of Myanmar and the anthropoid status of Amphipithecidae. Am. J. Phys. Anthropol. 44, 70e71. Brown, B., 1922. Letter from B. Brown to W. D. Matthew, Oct. 10, 1922. AMNH Vertebrate Paleontology Archives, Barnum Brown Papers, VPA 16a, 2:5, box 2, folder 5. Brown, B., Gregory, W.K., Hellman, M., 1924. On three incomplete anthropoid jaws from the Siwaliks, India. Am. Mus. Novit. 130, 1e8. Chaimanee, Y., Yamee, C., Tian, P., Chavasseau, O., Jaeger, J.-J., 2008. First middle Miocene sivaladapid primate from Thailand. J. Hum. Evol. 54, 434e443. Chopra, S.R.K., Vasishat, R.N., 1980. A new Mio-Pliocene Indraloris (Primate) material with comments on the taxonomic status of Sivanasua (Carnivore) from the Sivaliks of the Indian Subcontinent. J. Hum. Evol. 9, 129e132. Ciochon, R.L., Gunnell, G.F., 2002. Eocene primates from Myanmar: historical perspectives on the origin of Anthropoidea. Evol. Anthropol. 11, 156e168. Ciochon, R.L., Gingerich, P.D., Gunnell, G.F., Simons, E.L., 2001. Primate postcrania from the late middle Eocene of Myanmar. Proc. Natl. Acad. Sci. 98, 7672e7677. Colbert, E.H., 1935. Siwalik mammals in the American Museum of Natural History. Trans. Am. Phil. Soc. 26, 1e401. Conroy, G.C., 1987. Problems of body-weight estimation in fossil primates. Int. J. Primatol. 8, 115e137. Ducrocq, S., Jaeger, J.eJ., Chaimanee, Y., Suteethorn, V., 1995. New primate from the Palaeogene of Thailand and the biogeographical origin of anthropoids. J. Hum. Evol. 28, 477e485. Dunn, R.H., Rose, K.D., Rana, R.S., Kumar, K., Sahni, A., Smith, T., 2016. New euprimate postcrania from the early Eocene of Gujrat, India, and the strepsirrhinehaplorhine divergence. J. Hum. Evol. 99, 25e51. Fermor, L.L., 1945. Charles Stewart Middlemiss 1859e1945, 5. Obituary Notices of Fellows of the Royal Society, pp. 262e286. Fleagle, J.G., 2013. Primate Adaptation and Evolution, 3rd Edition. Academic Press, San Diego. Flynn, L.J., Morgan, M., 2005. New lower primates from the Miocene Siwaliks of Pakistan. In: Lieberman, D.E., Smith, R.J., Kelley, J. (Eds.), Interpreting the Past: Essays on Human, Primate and Mammal Evolution in Honor of David Pilbeam. Brill Academic Publishers, Boston, pp. 81e102.

Gaur, R., Chopra, S.R.K., 1983. Palaeoecology of the middle Miocene Siwalik sediments of a part of Jammu and Kashmir State, India. Palaeogeogr. Palaeoclimatol. Palaeoecol. 43, 313e327. Gebo, D.L., 2002. Adapiformes: phylogeny and adaptation. In: Hartwig, W.C. (Ed.), The Primate Fossil Record. Cambridge University Press, Cambridge, pp. 21e43. Gebo, D.L., Dagosto, M., Beard, K.C., Wang, J., 1999. A first metatarsal of Hoanghonius stehlini from the late middle Eocene of Shanxi Province, China. J. Hum. Evol. 37, 801e806. Gebo, D.L., Beard, K.C., Ni, X., Dagosto, M., 2015. Distal phalanges of Eosimias and Hoanghonius. J. Hum. Evol. 86, 92e98. Gilbert, C.C., 2005. Dietary ecospace and the diversity of Euprimates during the early and middle Eocene. Am. J. Phys. Anthropol. 126, 237e249. Gilbert, C.C., 2013. Cladistic analysis of extant and fossil African papionins using craniodental data. J. Hum. Evol. 64, 399e433. Gilbert, C.C., Maiolino, S.A., 2015. Comment to ‟Primates in the Eocene” by Gingerich (2012). Palaeobiol. Palaeoenv. 95, 237e241. Gilbert, C.C., Rossie, J.B., 2007. Congruence of molecules and morphology using a narrow allometric approach. Proc. Natl. Acad. Sci. 104, 11910e11914. Gilbert, C.C., Frost, S.R., Strait, D.S., 2009. Allometry, sexual dimorphism, and phylogeny: a cladistics analysis of extant African papionins using craniodental data. J. Hum. Evol. 57, 298e320. Gilbert, C.C., Patel, B.A., Friedman, A.C., Pugh, K.D., Fleagle, J.G., Patnaik, R., 2014. New Lower Siwalik localities near Ramnagar, India: Implication for the earliest Asian Great Apes and other mammalian lineages. Palaeontol. Soc. Ind. 5, 353e365. Gingerich, P.D., 1986. Early Eocene Cantius torresi-oldest primate of modern aspect from America. Nature 320, 319e321. Gingerich, P.D., 1995. Sexual dimorphism in earliest Eocene Cantius torresi (Mammalia, Primates, Adapoidea). Contributions from the Museum of Paleontology, the University of Michigan 29, pp. 185e199. Gingerich, P.D., Sahni, A., 1979. Indraloris and Sivaladapis: Miocene adapid primates from the Siwaliks of India and Pakistan. Nature 279, 415e416. Gingerich, P.D., Sahni, A., 1984. Dentition of Sivaladapis nagrii (Adapidae) from the late Miocene of India. Int. J. Primatol. 5, 63e79. Gingerich, P.D., Holroyd, P.A., Ciochon, R.L., 1994. Rencunius zhoui, a new primate from the late middle Eocene of Henan, China, and a comparison with some early Anthropoidea. In: Fleagle, J.G., Kay, R.F. (Eds.), Anthropoid Origins. Plenum Press, New York, pp. 163e177.
e res de Rians (Eoc ne Infe
rieur, Provence). PalaeGodinot, M., 1981. Les mammife overtebrata 10, 43e126. Godinot, M., 1988a. Le gisement du Bretou (phosphorites du Quercy, Tarn et Gar
e
bre
s de l'Eoc ne supe
rieur. VI.dPrimates onne, France) ets a faune de verte Palaeontographica Stuttgart (A) 205, 113e127.
s de Bouxwiller (Eoce ne moyen, Alsace) et Godinot, M., 1988b. Les primates adapide  la compre
hension de la faune de Messel et a  l'e
volution des leurapport a Anchomomyini. Cour. Forsch.-Inst. Senckenberg 107, 383e407. Godinot, M., 2015. Fossil record of the Primates from the Paleocene to the Oligocene. In: Henke, W., Tattersall, I. (Eds.), Handbook of Paleoanthropology. Springer-Verlag, Berlin, pp. 1137e1259. Gregory, W.K., Hellman, M., Lewis, G.E., 1938. Fossil anthropoids of the YaleCambridge India expedition of 1935. Carnegie Inst. Washington Publ. 495, pp. 1e27. Gupta, S.S., 1997. Study and documentation of vertebrate fossils from the Siwalik Group of Jammu Sub-Himalayan foot hills. Geol. Surv. Ind. 129, 5e7. Gupta, S.S., 2000. Lithostratigraphy and structure of the Siwalik succession and its relationship with the Murree succession around Ramnagar area, Udhampur District, J and K. Himal. Geol. 21, 53e61. Gupta, S.S., Verma, B.C., 1988. Stratigraphy and vertebrate fauna of the Siwalik group, Mansar-Uttarbani section, Jammu district, Jammu and Kashmir. J. Palaeontol. Soc. Ind. 33, 117e124. Hammer, Ø., Harper, D.A.T., Ryan, P.D., 2001. PAST: Paleontological Statistics software package for education and data analysis. Palaeontol. Electron. 4, 9. Kelley, J., 2002. The hominoid radiation in Asia. In: Hartwig, W.C. (Ed.), The Primate Fossil Record. Cambridge University Press, Cambridge, pp. 369e384. Kelley, J., 2005. Twenty-five years contemplating Sivapithecus taxonomy. In: Lieberman, D.E., Smith, R.J., Kelley, J. (Eds.), Interpreting the Past: Essays on Human, Primate and Mammal Evolution in Honor of David Pilbeam. Brill Academic Publishers, Boston, pp. 123e143. € €nen Olschiefer Koenigswald, W.V., 1979. Ein Lemurenrest aus dem eoza der Grube €ontol. Zeitschrift 53, 63e76. Messel bei Darmstadt. Pala Koenigswald, W.V., Habersetzer, J., Gingerich, P.D., 2012. Pedal distal phalanges of the Eocene adapoids Europolemur and Darwinius compared to phalanges of Notharctus and other primates. In: Lehmann, T., Schaal, S.F.K. (Eds.), Messel and the Terrestrial Eocene. Proceedings of the 22nd Senckenberg Conference. Palaeobio. Palaeoenv. 92, 539e565. Lewis, P.O., 2001. A likelihood approach to estimating phylogeny from discrete morphological character data. Syst. Biol. 50, 913e925. Maiolino, S., Boyer, D.M., Bloch, J.I., Gilbert, C.C., Goenke, J., 2012. Evidence for a grooming claw in a North American adapiform primate: Implications for anthropoid origins. PLoS One 7 e29135, 1e28.
, J., Roig, I., Seiffert, E.R., Sola, S.-M., Boyer, D.M., 2016. Astragalar and Marigo calcaneal morphology of the middle Eocene primate Anchomomys frontanyensis (Anchomomyini): Implication for early primate evolution. J. Hum. Evol. 91, 122e143.

C.C. Gilbert et al. / Journal of Human Evolution 102 (2017) 21e41 Marivaux, L., Welcomme, J.eJ., Ducrocq, S., Jaeger, J.eJ, 2002. Oligocene sivaladapid primate from the Bugti Hills (Balochistan, Pakistan) bridge the gap between Eocene and Miocene adapiform communities in southern Asia. J. Hum. Evol. 42, 379e388. Marivaux, L., Beard, K.C., Chaimanee, Y., Jaeger, J.eJ., Marandat, B., Soe, A.N., Tun, S.T., Kyaw, A.A., 2008. Proximal femoral anatomy of a sivaladapid primate from the late middle Eocene Pondaung Formation (Central Myanmar). Am. J. Phys. Anthropol. 137, 263e273. Marivaux, L., Ramdarshan, A., Essid, E.M., Marzougui, W., Ammar, H.K., Lebrun, R., Marandat, B., Merzeraud, G., Tabuce, R., Vianey-Liaud, M., 2013. Djebelemur, a tiny pre-tooth-combed primate from the Eocene of Tunisia: a glimpse into the origin of crown strepsirrhines. PLoS One 8, e80778. http://dx.doi.org/10.1371/ journal.pone.0080778. Nanda, A.C., Sehgal, R.K., 1993. Siwalik mammalian faunas from Ramnagar (J and K) and Nurpur (H.P.) and lower limit of Hipparion. J. Geol. Soc. Ind. 42, 115e134. Ni, X., Li, Q., Li, L., Beard, K.C., 2016. Oligocene primates from China reveal divergence between African and Asian primate evolution. Science 352, 673e677. Pan, Y., 1988. Small fossil primates from Lufeng, a latest Miocene site in Yunnan Province, China. J. Hum. Evol. 17, 359e366. Pan, Y., Wu, R., 1986. A new species of Sinoadapis from the Lufeng hominoid locality. Acta Anthropol. Sinica 5, 31e40. Parmar, V., Prasad, G.V.R., 2006. Middle Miocene rhizomyid rodent (Mammalia) from the Lower Siwalik Subgroup of Ramnagar, Udhampur District, Jammu and Kashmir, India. Neu. Jah. Geol. Palaontol. 6, 371e384. Parmar, V., Magotra, R., Norboo, R., Veera, G.V.R., 2016. Rodent-based age appraisal of the Lower Siwalik Subgroup of Kalaunta, Ramnagar, Jammu, India. Alcheringa. http://dx.doi.org/10.1080/03115518.2016.1196435. Patel, B.A., Seiffert, E.R., Boyer, D.M., Jacobs, R.L., St Clair, E.M., Simons, E.L., 2012. New primate first metatarsals from the Paleogene of Egypt and the origin of the anthropoid big toe. J. Hum. Evol. 63, 99e120. Patnaik, R., 2013. Indian Neogene Siwalik mammalian biostratigraphy: an overview. In: Wang, X., Flynn, L.J., Fortelius, M. (Eds.), Fossil Mammals of Asia: Neogene Biostratigraphy and Chronology. Columbia University Press, New York, pp. 423e444. Patnaik, R., Cerling, T.E., Uno, K.T., Fleagle, J.G., 2014. Diet and habitat of Siwalk primates Indopithecus, Sivaladapis, and Theropithecus. Annls. Zool. Fennici. 51, 123e142. Pilgrim, G.E., 1927. A Sivapithecus palate and other primate fossils from India. Mem. Geol. Soc. Ind. 14, 1e26. Qi, T., Beard, K.C., 1998. Late Eocene sivaladapid primate from Guangxi Zhuang Autonomous Region, People's Republic of China. J. Hum. Evol. 35, 211e220. Ronquist, F., Teslenko, M., van der Mark, P., Ayres, D.L., Darling, A., Hohna, S., Larget, B., Liu, L., Suchard, M.A., Huelsenbeck, J.P., 2012. MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Syst. Biol. 61, 539e542. Rose, K.D., Rana, R.S., Sahni, A., Kumar, K., Missiaen, P., Singh, L., Smith, T., 2009. Early Eocene primates from Gujarat, India. J. Hum. Evol. 56, 366e404.

41

Rose, K.D., Holbrook, L.T., Rana, R.S., Kumar, K., Jones, K.E., Ahrens, H.E., Missiaen, P., Sahni, A., Smith, T., 2014. Early Eocene fossils suggest that the mammalian order Perissodactyla originated in India. Nature Communications 5. http://dx.doi.org/ 10.1038/ncomms6570. Sallam, H.M., Seiffert, E.R., 2016. New phiomorph rodents from the latest Eocene of Egypt, and the impact of Bayesian “clock”-based phylogenetic methods on estimates of basal hystricognath relationships and biochronology. PeerJ 4, e1717. http://dx.doi.org/10.7717/peerj.1717. Sehgal, R.K., Patnaik, R., 2012. New muroid rodent and Sivapithecus dental remains from the Lower Siwalik deposits of Ramnagar (J and K, India): Age implication. Quatern. Int. 269, 69e73. Seiffert, E.R., Simons, E.L., Clyde, W.C., Rossie, J.B., Attia, Y., Bown, T.M., Chatrah, P., Methison, M., 2005. Basal anthropoids from Egypt and the antiquity of Africa's higher primate radiation. Science 310, 300e304. Seiffert, E.R., Perry, J.M.G., Simons, E.L., Boyer, D.M., 2009. Convergent evolution of anthropoid-like adaptations in Eocene adapiform primates. Nature 461, 1118e1121. Seiffert, E.R., Costeur, L., Boyer, D.M., 2015. Primate tarsal bones from Egerkingen, Switzerland, attributable to the middle Eocene adapiform Caenopithecus lemuroides. PeerJ 3, e1036. Strait, D.S., Grine, F.E., Moniz, M.A., 1997. A reappraisal of early hominid phylogeny. J. Hum. Evol. 32, 17e82. Swofford, D.L., 2003. PAUP*. Phylogenetic Analysis Using Parsimony (* and Other Methods). Version 4.10b. Sinauer Associates, Sunderland, Massachusetts. Thiele, K., 1993. The holy grail of the perfect character: the cladistic treatment of morphometric data. Cladistics 9, 275e304.
couverte d'un Primate Adapiforme (Sivaladapinae Thomas, H., Verma, S.N., 1979. De ne moyen des Siwaliks de la re
gion de Ramnagar sub. fam. nov.) dans le Mioce (Jammu et Cachemire, Inde). CR Acad. Sci. 289, 833e836. Tong, Y.-S., Wang, J.-W., Huang, X.-S., 1999. Discovery of a nearly complete lower jaw of Hoanghonius stehlini Zdansky, 1930 (Mammalia, Primates). Vertebrata PalAsiatica 4, 105e119. Vasishat, R.N., 1985. Antecedents of Early Man in Northwestern India: Palaeontological and Palaeoecological Evidences. Inter-India Publications, New Delhi. Vasishat, R.N., Gaur, R., Chopra, S.R.K., 1978. Geology, fauna and palaeoenvironment of Lower Sivalik deposits around Ramnagar, India. Nature 275, 736e737. White, J.L., 2006. Functional morphology and evolution of the adapiform dentition, with particular emphasis on the Asian Sivaladapidae. Ph.D. Thesis, University of Iowa. Wiens, J.J., 2000. Coding morphological variation within species and higher taxa for phylogenetic analysis. In: Wiens, J.J. (Ed.), Phylogenetic Analysis of Morphological Data. Smithsonian Institution Press, Washington, pp. 115e145. Woo, J.K., Chow, M.C., 1957. New materials of the earliest primate known in China. Vert. PalAsiatica 1, 267e272. Wu, R., Pan, Y., 1985. A new adapid genus from the Miocene of Lufeng. Acta Anthropol. Sinica 4, 1e6.