early hominin biogeography in island southeast asia · early hominin biogeography in island...

ARTICLE

Early Hominin Biogeography in IslandSoutheast AsiaROY LARICK AND RUSSELL L. CIOCHON

Island Southeast Asia covers Eurasia’s tropical expanse of continental shelfand active subduction zones. Cutting between island landmasses, Wallace’sLine separates Sunda and the Eastern Island Arc (the Arc) into distinct tectonicand faunal provinces. West of the line, on Sunda, Java Island yields many fossilsof Homo erectus. East of the line, on the Arc, Flores Island provides one skele-ton and isolated remains of Homo floresiensis. Luzon Island in the Philippineshas another fossil hominin. Sulawesi preserves early hominin archeology. Thisinsular divergence sets up a unique regional context for early hominin dispersal,isolation, and extinction. The evidence is reviewed across three Pleistocene cli-mate periods. Patterns are discussed in relation to the pulse of global sea-levelshifts, as well as regional geo-tectonics, catastrophes, stegodon dispersal, andpaleogenomics. Several patterns imply evolutionary processes typical of oceanicislands. Early hominins apparently responded to changing island conditions fora million-and-a-half years, likely becoming extinct during the period in whichHomo sapiens colonized the region.

In 1859, Alfred Russell Wallace

identified two faunal provinces

within Island Southeast Asia (ISEA),

Sunda and the Arc. Wallace’s Linerepresents a series of sea-channelbarriers to the dispersal of largemammals between them. The provin-ces are based primarily on the conti-nental origin of large terrestrialmammals (Fig. 1, Box 1).1 West ofWallace’s Line, mammalian specieshave Eurasian origins; east of theline, “Wallacean” mammals andother vertebrates show a mixture ofEurasian and Australian origins.Similar, if less provincial differentia-tion can be observed for some spe-cies of fish, insects, and birds. Theline should pose a factor for ISEAearly hominin dispersal, isolation,and extinction.

EARLY HOMININBIOGEOGRAPHY IN ISLAND

SOUTHEAST ASIA

In 1891, Eugene Dubois’ namedPithecanthropus erectus (now Homoerectus) based on a calotte and femurfound at Trinil, in the Solo Basin ofeastern Java.2 Sangiran, also in the

Solo Basin, has since produced more

than 80 Homo erectus cranial anddental fossils. The Sangiran and Tri-

nil fossils have thick cranial vaults

and cranial capacities of 840 to 1,059cc.3 A much later set of Solo Basin

Homo erectus fossils, from Ngandongand related sites, have cranial

capacities reaching 1,250 cc.In 2003, at Liang Bua, on Flores,

east of Wallace’s Line, Homo flore-

siensis was defined on the basis of

one nearly complete skeleton and

fragmentary remains of several indi-

viduals.4,5 Compared to Sunda Homo

erectus, the fossils from Liang Bua

have a very small cranial capacity

(417 cm3). In relation to most Pleis-

tocene early hominins, the Liang

Bua skeleton is short (1.06 m) and

has primitive wrists and large feet,

as well as a late age (�60 ka).6 Other

members of the Liang Bua vertebrate

fauna share similar insular charac-

ters (Box 2).6 When compared with

related species on other ISEA land-

masses, the Liang Bua fauna show

signs of isolation on Flores for a sig-

nificant part of the Pleistocene.7

In 2007, fragmentary fossils werecollected from Callao Cave on the

island of Luzon, in the Philippines.8

Because of its small dimensions andgracile morphology, a complete meta-

tarsal resembles those in small-bodied early hominins, including

Homo habilis and Homo floresiensis.Its date, 66.7 ka, is close to that of

Liang Bua. In 2014, an upper molartooth row of archaic character was

recovered, as were additional smalllimb bones.9 The new Callao finds

suggest a possible third group to the

ISEA Pleistocene hominin population

Roy Larick owns Bluestone Heights, anenvironmental education and consultingfirm in Cleveland, OH. He is a Paleolithicarcheologist who has done field workcovering Europe, Africa, and Asia. He is afounding member of the Iowa-BandungJava Project. Email: [email protected]

Russell Ciochon is Professor of Anthro-pology and Director of Museum Studiesat the University of Iowa, Iowa City. He isa paleoanthropologist with field workprojects throughout Asia. He also is afounding member of the Iowa-BandungJava Project. Larick and Ciochon havecollaborated on paleoanthropologicalprojects in Vietnam, China, and Indone-sia. Email: [email protected]

Key words: Sunda; insular dwarfism; island rule;

Homo erectus; Homo floresiensis; Marine Isotope

Stages (MIS); Java; Flores; Luzon; Sulawesi;

Timor

VC 2015 Wiley Periodicals, Inc.DOI: 10.1002/evan.21460Published online in Wiley Online Library(wileyonlinelibrary.com).

Evolutionary Anthropology 24:185–213 (2015)

and a second colonization east ofWallace’s Line.

Early hominins are now definitivelysituated on Sunda and at two wide-spread points on the Arc. Both Arcfossil hominins have the small sizeand specialized skeletal traits seen ininsular evolutionary contexts and, toa lesser extent, in the earliest Homoerectus from Eurasia (Dmanisi). TheArc hominins diverge greatly from allknown Sunda forms. After more thana century of accumulating fragmen-tary evidence, ISEA early homininbiogeography is now a significantresearch topic. Here we review rele-vant evidence from Sunda and theArc within a framework of Pleisto-cene climate change, our primarygoal being to evaluate potential rolesfor well-known insular evolutionaryprocesses in ISEA early hominin evo-lution (Box 2).

While ISEA early hominin fossilsare few and spatially isolated,Pleistocene-age stone artifacts are

numerous and widespread. Archeol-ogy therefore fleshes out early homi-nin biogeography. The historicaltrend of archeological research par-allels that of fossils. Recent researchprovides a richer comparative base(Box 3). Stone tools are incorporatedinto the review when an excavatedstratigraphic sequence containsfauna and artifacts within a recog-nized geological level and when asequence-long sampling strategy hasconsistently produced Pleistoceneages.

Using dated Marine Isotope Stages(MIS) global events and a long chro-nology for regional occurrences, wecan begin to comprehend ISEA bio-geographic events in their globalcontexts (Fig. 2, Box 4). Three globalevents can be tied to crucial regionalbiogeographic transitions: the Oldu-vai paleomagnetic event, the LateEarly Pleistocene Revolution (akaMid-Pleistocene Revolution), and theMid-Brunhes Event. This review has

three sections corresponding to theseevents. A period framework makesfor some repetition in presentingsites with long stratigraphic sequen-ces. Nevertheless, parsing regionaldevelopments by period helps toidentify the effects of climate changeand several regional environmentalcatastrophes (Box 5). Figure 2presents the overall scheme. Eachperiod section has a synoptic tablefor the relevant events, sites, lithics,and fauna. Table 1, for example,presents the earlier Pleistocenebenchmarks.

OLDUVAI SUBCHRON, OREARLIER PLEISTOCENE

During the earlier Pleistocene(�2.6-0.9 Ma), a 41-kyr orbital cycledrove global climate. This periodic-ity, based on earth’s orbital obliquity,exemplifies one of three such orbitalpatterns, known as Milankovitchcycles.10 During this time, glacial-

Figure 1. Southeast Asia: sedimentary basins and catastrophe origin points of relevance to early hominin paleoanthropology. Terrestrial sur-face extends to 100 m below current sea level. We followed Huxley’s modification of Wallace’s Line208 as illustrated in Cooper and Stringer204.Boxes show areas detailed in Figure 3. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

186 Larick and Ciochon ARTICLE

http://wileyonlinelibrary.com

interglacial cycles were relativelyshort and had low temperatureamplitude.11,12 Toward the middle ofthis phase, the Olduvai Subchronmarks a brief geomagnetic polereversal event, 1.98-1.79 Ma.13 Soonafter Olduvai, increases in monsoonintensity were recorded in the terres-trial contexts of the Lake Turkana(Kenya) and Heqin (China)basins.14,15 Glacial-interglacialcycling had an increased effect on

the size and distribution of Sundalandmasses. Northern Hemisphereglacial ice advanced significantlybetween about 1.8 and 1.74 Ma (MIS62, 60, and 58), about 1.56 Ma (MIS52), between 1.24 and 1.1 Ma (MIS36, 34, and 30), and about 0.9 Ma(MIS 22).16 During these stages, sea-level lowstands opened emergentlandmasses throughout ISEA.17,18 Itwas, apparently, just after the Oldu-vai event that Pleistocene Eurasian

mammals arrived on an emergent

Sunda.19–21

Sunda

Sangiran, Java (Indonesia)

Much of the Solo Basin lies at

about 78 S latitude and sits above theIndonesian subduction zone (Fig.3A). Mammalian fossils are pre-served within a 500,000-year

Figure 2. ISEA Pleistocene chronology and site correlation. Even-numbered MIS represent cooler phases (Northern hemispheric glacials);odd-numbered stages represent warmer phases (Northern hemispheric interglacials). MIS ages are drawn from Bowen and Sikes,272

Berger et al.,109 and Willoughby.273 Ages before MIS 19 are approximate. Since early hominin fossils are known only from sites in Javaand Liang Bua, Flores, artifacts serve as the evidence of early hominins at other sites. FAD 5 first appearance datum; LAD 5 lastappearance datum. Except where noted, site or event placement marks earliest occurrence. Blue lines indicate first-occurrence pre-sumed hominin continuous occupation in each regional site sequence unless otherwise noted. Sites with named formations and/orgeologic designations: Lower Lahar, Lower Lahar Unit (Sangiran Formation); Sangiran, Sangiran Formation; Bapang, Bapang Forma-tion; Ngebung, Bapang Formation; Song Terus, karst cave; Ngandong, 20 m terrace of Solo River; Wolo Sege, Tangi Talo, MataMenge, Boa Lesa, and Kobatuwa, Ola Kile Formation; Liang Bua, karst cave; S Enrile Q, Awidon Mesa Formation; Cagayan, Ilaganand Awidon Mesa Formations; Arubo, Sierra Madre foothills; Callao, karst cave; Cabenge, Walanae Basin fill. [Color figure can beviewed in the online issue, which is available at wileyonlinelibrary.com.]

ARTICLE Early Hominin Biogeography in Island Southeast Asia 187


sequence (�1.53-0.98 Ma) of aggrad-ing volcaniclastic sediments.19 Cur-rently, more than 80 fossil fragmentsof Homo erectus are known fromSangiran (Fig. 4). The area has threeprimary sedimentary deposits: theLower Lahar Unit, the Sangiran For-mation, and the Bapang Formation.

Lower Lahar Unit. As early as 1.90Ma, a nearby volcanic cone, pro-duced a massive lahar-type debrisflow. Sangiran represents a relativelydistal point of the flow, where thelahar entered marine conditions onSunda’s south coast. Incorporatedfossils indicate that the lahar pushedthrough numerous environmentsbefore arriving at Sangiran. A cervid

mandibular fragment reflects terres-trial conditions in the higherreaches.20 Freshwater mollusks indi-cate swamps or shallow lakes in thelower reaches.22 Sangiran itselfreflects near-shore marine environ-ments transformed into estuarineand marsh settings. Shortly there-after, glacio-eustatic sea level regres-sions exposed lahar-infilled lagoonsand near-shore environments to cre-ate terrestrial habitats.23 Since nei-ther hominin fossils nor stoneartifacts were incorporated, it isassumed that the Lower Lahar Unitpredates the arrival of Homo erectus.

Sangiran Formation. A sequence ofdark-colored lacustrine siltstones,

mudstones, and thin tuffs overlies theLower Lahar Unit. These representslow-moving streams draining nearbylow volcanic highlands into shallowlakes. Such watercourses intermit-tently flooded lake margins andmarshes, transforming coastal lagoonsinto inland lakes. Occasional volcaniceruptions deposited thin blankets ofash. Lake-edge and marsh environ-ments supported sedges, ferns, water-tolerant grasses, and trees.24,25 Theassociated fauna comprised aquaticand semi-aquatic vertebrates suchpygmy hippo (Hexaprotodon), croco-dile (Crocodylus), and tortoise (Geo-chelone), as well as turtle and fish(fragmentary remains).21,26–28 Wetgrasslands with scattered shrubs

Box 1. Geomorphologic Setting

ISEA encompasses Sunda andthe Arc. ISEA is the result of theIndian and Pacific oceanic platessubducting under the Eurasian ter-restrial plate. Regional topographyranges from broad coastal plainand continental shelf to volcanicislands, plateaus, and deep seatrenches (Figs. 1 and 8).

Sunda: Much of the IndonesianArchipelago lies on the Sundashelf, a vast, now mostly sub-merged southward extension of theEurasian continental plate.206,207

The continental shelf extendssouthward from the present South-east Asia mainland toward Javaand the Indian Ocean (8.18 S).Here, the term “Sunda shelf” isreserved for the continental projec-tion. while “Sunda” refers to theshelf’s island landmasses. The term“Sunda” updates the nineteenth-century names of “Sundaland” and“Sunda Land.” Today, Sunda takesin the current large landmasses ofthe Malay Peninsula, Borneo,Sumatra, and Java, as well assmaller islands such as Bali.

Eastern Island Arc: East ofSunda, the sea bottom is complexwith trenches and ranges. Thelandmasses of this area constitutethe Arc, which stretches from

Luzon in the north, southwardthrough Sulawesi and the MalukuIslands and, in the extreme south, toFlores, Sumba, Timor, and relatedsmall islands. Much of the Arc liesin Wallacea, the geographic and eco-logical transition zone betweenSunda and Sahul (Australia andNew Guinea).208 During the Plioceneand Pleistocene, especially duringperiods of glacio-eustatic sea-leveldrawdown, large mammals, includ-ing primates, evidently penetratedcurrent marine barriers betweenSunda and Wallacea.209,210 Fossiland archeological evidence indicatesthat early hominins traversed cur-rent straits numerous times duringthe Pleistocene.

Eustasy, Tectonics, Volcanism:Glacio-eustatic sea level changes of�125 m are documented for thePleistocene.211 Glacial eustasyrepeatedly redistributed habitableisland landmasses and interislandconnections. Glacial period sea-level lowstands maximized condi-tions for regional dispersals. Inter-glacial highstands separated earlyhominin groups and set up condi-tions for insular endemism. Severalsubduction zones have made fordeep sea barriers of full Pleistoceneduration. Trending north-south to

create Wallace’s line are the Mind-oro Strait, Sulu and Celebes Seas,and the Makassar and LombokStraits. Trending east-west are theFlores and Banda Seas and, southof Flores, the Savu Sea.

Regional tectonic events mayhave been involved with specificand short-lived dispersal pathways.Sartono implicated small-scaleuplift in relation to dispersal corri-dors along the Palawan, Sulu, San-gihe, and Selayar archipelagos (Fig.8).85 While this particular model isnow dated, small-scale tectoniceffects could be approached in newterms. Within the area covering theSeas of Java, Banda, and Molucca(Fig. 8), there are numerous plate-lets with active convergent anddivergent boundaries.212 Such pla-telets are subject to forming pop-up blocks.213 This process couldhave enhanced the formation oflowstand land bridges. Regionaltephra catastrophes could rear-range local habitats. The Sundasedimentary record shows frequentvolcanic eruptions along plate mar-gins after the Olduvai Subchron.Truly large emissions could extendlocal landmasses seaward, provid-ing larger lowland areas and, insome cases, small land bridges.214


occupied slightly higher landscapes.Still higher, better-drained parts of thelandscape supported a community ofsedges, grasses, ferns, and scatteredtrees.29 Fauna included stegodon(Stegodon elephantoides), cervids (Cer-vus zwanni, C. hippelaphus, and Cer-vus sp. indet.) and one small bovid(Duboisia santeng), as well as Homoerectus.19,26,27,30 The hominin-bearing

upper reaches of the formation date tobetween 1.66 and 1.57 Ma.23

At Bukuran, apparent technologi-cal cutmarks have been observed onbovid bones from legacy collections.Two specimens show marks outsidethe range of natural causes. Whilestone tools are absent from the San-giran Formation, molluscan shells

are abundant. In a series of experi-

ments, clamshell cutting tools best

replicated the Bukuran bovid bone

cut marks.31

Bapang Formation. Between 1.6 and1.5 Ma, volcanic cones grew north-west and southeast of Sangiran.Larger, more powerful streams beganscouring and infilling the local low-lands.23 A stark fluvial erosion

Figure 3. ISEA early hominin sites: A, central/eastern Java; B, central Flores; C, northern Luzon; D, southwest Sulawesi. [Color figure canbe viewed in the online issue, which is available at wileyonlinelibrary.com.]



surface often marks the contactbetween the Sangiran and BapangFormations. This contact represents aperiod of net sediment removal fromthe Sangiran area.29 Immediatelyabove the contact, the GrenzbankZone has poorly sorted coarse sedi-ments. Heavy clasts, including verte-brate fossils, can be highlyfragmented and indicate multiplereworking. Hominin fossils are foundthroughout the Bapang. However, incomparison with higher reaches ofthe formation, the Grenzbank lackshominin skeletal elements of rela-tively low density, such as calottes,and retains only the denser mandibu-lar and maxillary fragments andteeth. 40Ar/39Ar radiometric analysissuggests that the deposit began accu-mulating more than 1.5 Ma and con-tinued to 0.9 Ma,19 thus recordingclimate cycles in the range of MIS 47-23.

The Bapang sequence contains

numerous paleosols. These devel-oped on riverine landscapes, repre-

senting riparian forest, savanna, andopen woodland environments.29

Paleosol morphology and carbon iso-

tope values indicate a long-term shifttoward longer annual dry seasons.29

The more open habitat supportedcarnivores (Panthera), pigs (Sus bra-chygnatus), cervids (Axis lydekkeri),

large bovids (Bubalus palaeokerabau,Bibos palaesondaicus), stegodon(Stegodon trigonocephalus), and pri-mates (Homo erectus, Macaca sp.indet., Trachypithecus cf. aura-tus).27,32,33 After about 0.9 Ma, Poh-jajar Formation fluvial depositscovered the Bapang sequence with ahigher proportion of air-fall tuffs,fluvially reworked ash fall, and lahardeposits (Fig. 2). The Pohjajar hasnot yielded Homo erectus remains atSangiran or elsewhere.

Trinil, Java

In its middle course, the SoloRiver makes an abrupt northwardturn to cut through the KendengHills (Fig. 3A). In 1891, EugeneDubois made the Java Man discoveryof a calotte and femur on the SoloRiver bank at Trinil. Dubois namedthe collection deposit the Hauptkno-chenschicht (main bone layer) of Tri-nil.34,35 Much later, four morefemora were collected in the samearea.36–38 The Trinil skullcap lieswithin the range of morphologicalvariation for the large collection ofSangiran specimens. Together, theSangiran and Trinil crania shouldrepresent the earlier Sunda Homoerectus paleodeme.

The stratigraphic relationshipamong the fossils has always beenquestioned, especially between thecalotte and the original Femur I.39,40

Recently, all five Trinil femora havebeen morphologically compared andtheir structural and density charac-teristics evaluated by computedtomography.41 Femur I is anatomi-cally more modern and less fossil-ized than Femora II-V. Femur I isapparently younger than the calotte,while Femora II-V may be moreclosely related in time to the Homoerectus calotte.41

Trinil has seen little field investiga-tion since 1891. The stratigraphyremains unimproved and contentionpersists about the age of Hauptkno-chenschicht. The main bone layer isnow considered an overbank depositbuilt up during repeated floods. Itmay contain materials of differentorigins and ages.42,43 The Duboisfauna collection, curated at the ReiksMuseum in Leiden, has been inten-

sively studied. Old endemic speciesinclude Duboisia santeng and Axislydekkeri, as well as Stegodon trigono-cephalus, Bubalus palaeokerabau,Bibos palaesondaicus, Sus brachygna-thus, and Panthera tigris trinilen-sis.26,44 The assemblage has beencompared with that of the Bapangformation Grenzbank Zone, havingan age, using the short chronology,of around 1.0 Ma.26 The long chro-nology places the Grenzbank Zone at1.5 Ma (Box 4).19 In any event, com-parison with the Sangiran sequenceis of limited value.

Using Trinil and other legacy fau-nal collections, Storm modeled theecological role of ISEA Homo erec-tus.45 The number of identifiedspecimens and the minimum num-ber of individuals reflect trophic lev-els of primary and secondaryconsumers. Further, the numbers ofremains of Homo erectus, at sitessuch as Sangiran, resemble those oflarge carnivores. These numbers sug-gest that Homo erectus functioned asa carnivorous omnivore.45

A recent analysis of curated Haupt-knochenschicht materials is provoca-tive. Results suggest a significantlyyounger age for the site and thatHomo erectus made complex use offreshwater shellfish. 40Ar/39Ar andluminescence dating methods havebeen applied to sediment adhering tofreshwater mollusk shells (Pseudodon)in the surviving faunal collection.46

The results are indicative of the Mid-Brunhes Event (�480 ka). As the newage analysis depends on materialsremoved from stratigraphic context, itis difficult to make a definitive judg-ment. The dates and behavioralhypothesis are presented in the Mid-Brunhes Event section.

Eastern Island Arc

Soa Basin, Flores (Indonesia)

The Soa Basin covers 200 km2 ofwest-central Flores and containstwo volcaniclastic sedimentaryunits. At its base, the Ola Kile For-mation consists of andesitic brec-cias and volcanic mudflows withminor interbedded tuffaceous silt-stones, sandstones, and lava flows.47

From near the top of the formation,

Figure 4. Homo erectus cranium (S 17):Bapang Formation below the Middle Tuff,Sangiran, Solo Basin, Java. Cranialcapacity: 1,004 cc.274 (Photograph: R. Cio-chon). [Color figure can be viewed in theonline issue, which is available at wileyonli-nelibrary.com.]




a fission-track analysis provides aminimum age of 1.86 6 0.12 Ma.Above Ola Kile, the Ola Bula For-mation comprises 100 m of volcanicand fluvio-lacustrine deposits. Likethe Solo Basin Bapang Formation,the Ola Bula is highly volcaniclasticand indicates young riverine habi-tats and the presence of lakes. TwoSoa Basin sites, one archeologicaland one paleontological, revealFlores sedimentary and faunal envi-ronments toward the end of theEarly Pleistocene (Fig. 3B).

Tangi Talo. This paleontologicallocale lies low in the Ola Bula strati-graphic sequence. Fauna includedwarf Stegodon (Stegodon sondaari),giant tortoise (Geochelone), andKomodo dragon (Varanus komodoen-sis). There are no signs of butcheringand no stone artifacts. The TangiTalo small-bodied S. sondaari is theearliest stegodon on Flores. An over-lying tuff of volcanic debris is dated

to 0.90 6 0. 07 Ma.48 Dating analysescontinue at Tangi Talo, with newresults based on 40Ar/39Ar eruptionage suggesting 1.3 Ma.49

Wolo Sege. This archeological sitelies in the Ola Bula basal tuff interval,just above the Ola Kile breccias. Thesedimentary environment is highlyvolcaniclastic and represents a stagebefore the fluvial-lacustrine land-scapes had truly developed. WoloSege has in-situ stone artifacts,including some Acheulean-like imple-ments50 (Fig. 5). There is no in-situfauna. Overlying the artifact layers isan ignimbrite with an 40Ar/39Ar erup-tion date of 1.02 6 0.02 Ma.47

The Soa Basin’s Ola Bula depositsprovide the oldest record of Pleisto-cene fauna and stone tools east ofSunda. The Wolo Sege artifacts indi-cate that hominins arrived on Floreswell before 1.0 Ma. The young vol-canic environment represents theconditions that hominins and other

large mammals encountered on

entering Flores.47

Walanae Basin, South Sulawesi

(Indonesia)

Cabenge. The Walanae Basin lies on

Sulawesi’s southwest peninsula (Fig.

3D). In the late 1940s, H. R. van

Heekeren collected surface stone

artifacts and fauna near the village

of Cabenge (formerly Tjabenge). The

Walanae fauna included giant

tortoise (Testudo morgae), pygmy

elephant (Archidiskodon celebensis),

stegodon (Stegodon sompoensis),

giant pig (Celebochoerus heekereni),and dwarf buffalo (Anoa depressicor-nis).51 Although the presence of

Archidiskodon suggested an earlier

Pleistocene age, Van Heekeren linked

the Tjabenge stone tool industry

with this fauna.52 He cautioned,

nevertheless, that the Walanae Archi-diskodon could represent a Middle

Pleistocene descendant of a long-

Box 2. Insular Evolutionary Process

Oceanic islands are isolated set-tings in which natural selectionand genetic drift can intensify andthus accelerate evolutionary rates.Immigrating island populations aresmaller and less genetically diversethan the mainland source popula-tion, resulting in a founder effect.Islands have lower biodiversitythan do mainland areas, whichresults in species expanding andshifting ecological niches fromtheir mainland counterparts. More-over, island size is linked with fau-nal turnover, with smaller islandshaving higher extinction rates.215

These conditions force microevolu-tionary changes that may lead tomacroevolutionary changes, includ-ing speciation. Van Valen’s “islandrule” observes that in long-termisland contexts, large-bodied mam-mals tend to become smaller andsmaller ones bigger.216 The effectsof the island rule tend to be inver-sely proportional to an island’s

size217 and positively correlatedwith its degree of isolation fromthe mainland.218 The smaller andmore isolated an island, the moresignificant the role of island rule.

Island dwarfing has been observedin a wide variety of both living andfossil mammals. There are well-known cases of Pleistocene probosci-dean dwarfing relating to Pleistocenesea-level fluctuations. These includeelephants on several Mediterraneanislands, mammoths on the Califor-nia Channel Islands, and stegodonsin Island Southeast Asia.

The strong hypothesis for islanddwarfing revolves around popula-tion size versus food availability.219

In general, island landmasses offerrelatively reduced nutritionalresources to relatively small popu-lations. Islands also tend to havefewer large predators. In generalresponse, mammals have feweryoung, while patterns of body sizeevolution are much accelerated.220

This is because large-bodied indi-viduals use more resources, soisland natural selection favorssmaller individuals which, overtime, produce smaller-bodied popu-lations.221 Smaller body size aids inmaintaining relatively large popula-tions on island resource bases.222

Pleistocene ISEA demonstratesnumerous cases of island endem-ism. At least five dwarfed species ofStegodon emerged in the region.Van Valen’s island rule is directlyapplicable in the large and smallforms observed in the Liang Buavertebrate fauna.6,216 These includegiant tortoise (Geochelone), giantrat (Spelaeomys or Hooijeromisnusatenggara), a very small stego-don (Stegodon florensis insularis),and the dwarfed hominin Homofloresiensis (Box 6). As a dwarfhominin, Liang Bua Homo flore-siensis fits well within itsISEA ecological and evolutionarycontext.6


lived endemic island form. Questionsarose about the association of theindustry with the fauna and the ageof each.53–56 With new research, theWalanae terraces are now inter-preted as the upper part of a normalbasin fill sequence uplifted duringthe Late Pleistocene.32 New work atnearby Talepu suggests that someartifact horizons have an early Pleis-tocene age.57

The Walanae basin volcaniclasticgravel series displays clasts of yellowchalcedony and red jasper. Theseand other highly colored, fine-grained siliceous rocks are the rawmaterials for the Tjabenge Indus-try.52 Tjabenge flakes are relativelysmall and thick, with signs of havingbeen struck in all directions fromirregular cores. Tool types includepoints, concave scrapers, core andkeeled scrapers, endscraper picks,and chopping tools. Van Heekerennoted similarities between the Tja-benge and Sangiran flake industries,including the use of small, highlycolored, fine-grain raw materials andthe production of irregular cores.He concluded that these and othersimilar flake industries were pro-duced by a single species of homi-nin52 (Fig. 5).

LATE EARLY PLEISTOCENEREVOLUTION OR LATE EARLY-EARLY MIDDLE PLEISTOCENE

The MIS 24-22 complex (�1.0-0.9Ma) marks a shift in glacial-interglacial forcing from the 41-kyrobliquity-based cycle to a 100-kyreccentricity-based cycle.15 The MIS24-22 complex has thus been calledthe late Early Pleistocene Revolution(EPR) or the Mid-Pleistocene Revolu-tion.58–61 The full period of transi-tion, 1.25 Ma to 700 ka, has beentermed the Mid-Pleistocene Transi-tion.62–64 We call it the “late Early-early Middle Pleistocene.”

After EPR, glacial periods becamelonger and more evenly distributed.Longer glacials served to increaseglobal ice volume, thereby pushingsea-level lowstands as much as 50 mlower than during the earlier Pleisto-cene.62 Short but warm interglacialsdecreased ice volume quickly. The

TAB

LE1.

Early

Ple

isto

ce

ne

Syn

op

ticC

ha

rt

MIS

sta

rt

�k

a

en

d

�k

ae

ve

nts

ag

e

ka

Java

Flo

res

Sula

we

si

tre

nd

ssi

tes

lith

ics

Ste

go

do

nsi

tes

lith

ics

Ste

go

do

nsi

tes

lith

ics

Ste

go

do

n

24

975

920

EP

R/

Jara

mill

o990

Soa

eru

ptio

ns

Ca

be

ng

esm

all

flake

(Tja

be

ng

e)

S.so

mp

oe

nsi

sla

rge

rN

he

mis

he

reic

evo

lum

es

31

1070

Trin

ilS.

trig

on

oc

ep

ha

lus

Wo

loSe

ge

rad

ial

co

res

36

1200

sig

nifi

ca

nt

gla

cia

tio

n/

dra

w-d

ow

n37

1240

46-3

81400

1250

Tan

gi

Talo

son

da

ari

47

1450

San

gira

ne

rup

tio

ns

Ba

pa

ng

sma

llfla

ke(S

an

gira

n)

S.tr

igo

no

ce

ph

alu

s

62-5

21700

1550

San

gira

nS.

ele

ph

an

toid

es

sig

nifi

ca

nt

gla

cia

tio

n/

dra

w-d

ow

n66

1980

1790

Old

uva

isu

bc

hro

nLo

we

rLa

ha

rO

laK

ilefm

Dm

an

isi

H.

ere

ctu

sb

eg

in41-k

yr

ob

liqu

ity

cyc

le:

gla

cia

lc

yc

les

of

sho

rtd

ura

tio

na

nd

low

tem

pe

ratu

rea

mp

litu

de


MIS 23 interglacial is thereforeimplicated in flooding insular land-masses.65 At Heqin (China), 0.92 Mamarks the transition to a long period(0.92 to 0.13 Ma) during which theIndian Summer Monsoon (ISM) wasstructured by both southern highand northern low pressure. Duringthis period, ISM maxima coincidedwith Northern Hemisphere ice vol-ume minima (interglacials) and Ant-arctic temperature maxima.15 EarlyMiddle Pleistocene glacial sea-level

lowstands likely made for extensivedispersal corridors. Likewise, earlyMiddle Pleistocene interglacial high-stands led to insular isolation andendemic evolution.

The early Middle Pleistocene mayalso be associated with generallylower sedimentation rates. In thebasins of Soa (Flores) and Cagayan(Luzon), the period immediately sur-rounding the EPR is well representedby open-air archeological and faunalsites, but evidence of hominin habi-

tation trails off rapidly thereafter.The EPR marks the last knownoccurrence of Homo erectus at San-giran. The early Middle Pleistocenesaw two region-wide catastrophes:the Australasian tektite impact andthe oldest Toba Tuff super-eruption(Box 5).

Two areas, Sangiran and Soa,have sedimentary sequences thatextend from the Early to the MiddlePleistocene. Earlier, we presentedEarly Pleistocene evidence. We now

Box 3. Stone and Shell Tool Technology

Pleistocene ISEA stone toolassemblages are of three types.

1. Flake Industries. Early sur-face collecting at several locationshave led to this category, in whichflake and cobble tools are generallyad hoc in flake production andbifacial reduction. Flake industrieshave relatively few biface andcleaver-type tools.54,55,75 Flakeindustries show neither chronologi-cal development nor regionaldifferentiation.

Of relevance are the flake indus-tries named after their modern set-tlements of Cabenge (Sulawesi),Kota Tampan (Malay Peninsula),and Ngandong and Pacitan (Java).Questions of artifact age and verac-ity plagued the early surface finds.Late twentieth-century examina-tions of terrace structure andassemblage technology generallyascribed these surface finds to thelate Pleistocene.53–55,75 Recentresearch, nevertheless, has recov-ered stone tools in association withfauna and, sometimes, datablematerial. Generally speaking, theearliest well-dated archeologicalassemblages on both sides of Wal-lace’s Line have ages greater than900 ka. This pattern holds forJava,66 Flores,47 and Sulawesi.53

The earliest stone-tool horizon onLuzon is about 800 ka.97,101

2. Large-Flake Acheulean(LFA). This category is defined bythe biface- and cleaver-rich site of

Gesher Benot Ya’aqov (GBY) inIsrael.73,74 The LFA chaine operatoirearrives at simple yet effective cuttingtools in a small number of well-structured steps. LFA assemblagesfeature large (>10 cm) flakes, eitherraw or minimally retouched as uni-facial tools and simple Acheulean-type bifaces (convergent tip) andcleavers (broad tip). The LFA isassociated with coarse-grain quartz-ite and related materials rather thanfine-grain flint. Kombewa flaketypes are common as cleaver blanks.

As a cultural force, the LFAappeared by 1 Ma at Olorgesailieand spread throughout Africa andbeyond. While timing is imprecise,the LFA arrived in the Levant, Ibe-ria, and India by �600 ka. After500 ka, only sub-Saharan Africaholds sure LFA sites.73,74 Recently,the LFA profile has been appliedto assemblages from Ngebung(Sangiran, Java) and GunungSewu. The attribution was basedon typological grounds and a per-ceived link between the Pinjorfauna (associated with LFA sitesin India) and the “Stegodon-Homoerectus fauna” of Sunda.223 Bifacesand cleavers from the Cagayanand Arubo basin sites (Luzon)have also been termed Acheuleanand even Large-Flake Acheu-lean.103,104 Similar tools fromWolo Sege (Soa Basin, Flores) arealso called Acheulean.78 The ques-tion remains of whether or notAcheulean-like tools and reduction

sequences represent a Far Eastmanifestation of the Acheuleancultural formation.

3. Radial-Core Reduction. Thisis present on Flores, particularlyat Mata Menge and Liang Bua.Lithic resources were brought tothe site in the form of flakeblanks. The Acheulean-like ele-ments of Wolo Sege are not pres-ent at Mata Menge, where themore complex tools show a reduc-tion sequence based on centripetalor radial removals on the blanks.4

A radial reduction sequence isfound again at Liang Bua, 50 kmto the west and �800 kyr later.

ISEA has few Pleistocene arche-ological sites in which complexstone-tool technology is associatedwith diverse occupation debris.Several explanations have beenoffered. Early hominins main-tained a vegetarian diet that didnot require complex stonetools.224 Early hominins usedmore readily available nonlithicraw material, such as bamboo225

and wood.226 Since the late 1990s,evidence has mounted to supportthe use shellfish for food andshells for tools. Shell archeologyhas been published for three sites:Bukuran and Trinil on Java, andTo’os on Timor.31,46,87,227 Individ-ually, each site has marginal evi-dence of shell use. When pooled, ahypothesis for early homininshellfish use can nevertheless beentertained.


address the Middle Pleistocenerecord for Sangrian and Soa, as wellas other areas. Table 2 shows theevents, sites, lithics, and fauna thatwere present during the early MiddlePleistocene.

Sunda

Ngebung (Sangiran)

On Sangiran’s northwest flank, theNgebung Hills lie some 3 km NNW ofthe Sangiran and Bapang Formation

type sites. The terrain is a dissectedescarpment ranging from Quaternaryterraces atop the Pohjajar Formationdown through the Bapang to the San-giran Formation. From 1989 to 1994,a French-Indonesian team excavated

Figure 5. ISEA Early Pleistocene lithic artifacts. Ngebung 2, Sangiran, Solo Basin, Java: A, retouched large flake; B, polyhedron (courtesy ofF. Semah).66 Cagayan Basin, Luzon, Philippines (photograph: R. Ciochon). Arubo, Luzon, Philippines: bifacially retouched large flake (cour-tesy of E. Z. Dizon).97 Wolo Sege, Soa Basin, Flores: small flakes (courtesy of A. Brumm). Cabenge, Walanae Basin, Sulawesi: small flaketools (modified after H. R. van Heekeren).52 [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]



TAB

LE2.

Mid

-Ple

isto

ce

ne

Tra

nsi

tion

Syn

op

ticC

ha

rt

MIS

sta

rt

�k

a

en

d

�k

ae

ve

nts

ag

e

ka

Java

Flo

res

Luzo

n/B

ose

tre

nd

ssi

tes

lith

ics

Ste

go

do

nsi

tes

lith

ics

Ste

go

do

nsi

tes

lith

ics

Ste

go

do

n

11

420

360

MB

E420

ISEA

shift

toc

ave

oc

cu

pa

tio

n/

pre

serv

atio

n12

480

420

13

520

480

Gu

nu

ng

Sew

uu

plif

t&

kars

tfo

rma

tio

n

tro

pic

al

rain

fore

stN

ew

Gu

ine

a&

Au

stra

lia14

580

520

15

630

580

info

llow

ing

inte

rgla

cia

ls:

shift

tom

ore

mo

de

rnfo

rest

sw

ith

red

uc

ed

div

ers

ity

16

680

630

Tib

et

Pla

tea

um

info

rest

19

800

680

Bru

nh

es-

Ma

tuya

ma

778

Aru

bo

LFA

?

20

800

800

Ko

ba

tua

rad

ial

co

res

Ca

ga

ya

nLF

AS.

luzo

ne

nsi

s

21

Old

est

Tob

a/

Tekt

ite

s

803

Bo

aLe

sara

dia

lc

ore

sB

ose

bifa

ce

s

22

880

870

Ma

taM

en

ge

rad

ial

co

res

S.flo

ren

sis

Tib

et

Pla

tea

ufr

igid

gla

cia

l23

880

EP

R880

Ng

eb

un

g?

LFA

?S.

trig

on

oc

ep

ha

lus

24

975

920

Jara

mill

o990

He

qu

ing

shift

toc

old


at Ngebung 2, where several meters ofBapang riverine sands and gravelsoverlie the Grenzbank Zone.66 Fivestratigraphic ensembles yielded stoneartifacts and a highly fragmentedfauna.66 40Ar/39Ar dates on tuffaceouslayer heavy minerals produced ages of�800 ka.67 Our own analysis gave anage of 900 ka.68

The 2,0001 fossil bones and teethinclude more than 800 mammal fos-sils assigned to species: large bovidscomprise 45%; smaller cervids, 18%;and stegodons 8%.69 Endemic spe-cies include stegodon (Stegodon trig-onocephalus) and bovid (Duboisiasanteng). Mainland forms includepygmy hippo (Hexaprotodon sivalen-sis), buffalo (Bubalus paleokerabau),rhino (Rhinoceros sondaicus), andpig (Sus brachygnathus).

Bone surfaces are highly eroded.There are no cutmarks present, butbovid long bones appear to havebeen intentionally broken anddefleshed. Some Stegodon tusks arebroken as if to create ivory flakes.69

The large mainland forms, Bubalus,Rhinoceros, and Sus, should repre-

sent migrants contemporary with theinitial Homo erectus dispersal. Mostof these foraged in open settings but,for rest, depended on closed habi-tats. The fauna may therefore reflecta forest fringe setting.70

Ngebung stone artifacts featurelarge flakes (including Kombewatype), polyhedrons, cores, cleavers,and hammer stones with little lithicdebitage (Fig. 5).71 Spheroids arerecorded from an erosion zone in alower stratigraphic unit (ensembleA); these could be more recent intru-sions.66 Ngebung flakes, tools, andspheroids are made on andesite peb-bles and cobbles,72 likely quarriedfrom Bapang-era erosion surfaces inthe Lower Lahar Unit.23

The Ngebung assemblage has twopossible origins (Box 3). On onehand, the biface and cleaver ele-ments recall the Large Flake Acheu-lean (LFA).73,74 Should these be trueLFA artifacts, Ngebung would beamong the oldest such sites outsideof Africa.71 On the other hand, theNgebung assemblage resembles ISEAflake industries defined on surface

collections, such as the Sangiran and

nearby Pacitanian industries.53–55,75

It is possible that some or all of the

stone artifacts at Ngebung are sur-

face intrusions from a later period.

Eastern Island Arc

Soa Basin, Flores (Indonesia)

About 0.9 Ma, local volcanic erup-

tions contributed heavily to Ola Bula

basin infill. The earlier volcanic tuff-

dominated facies grades upwards

into a sandy interval dominated by

fluvio-lacustrine tuffaceous sand

layers.47 At this time, the earlier

Pleistocene dwarf stegodon (Stego-don sondaari) of Tangi Talo is not

represented, but appears to have

been replaced by a large-bodied form

(Stegodon florensis).76,77 S. florensislikely dispersed from a larger land-

mass during MIS 22, the first

extreme glacial period of the Pleisto-

cene. The accompanying sea-level

lowstand certainly opened dispersal

routes from Sunda to the Arc and

across the Arc. In such conditions,

Box 4. Long Chronology

During 125 years of ISEA paleonto-logical research, an early homininchronology has been difficult toresolve. The most productiveresearch area has been Sangiran, inCentral Java, Indonesia. With a thickvolcaniclastic sedimentary sequencepresenting at many locales, Sangiranhas produced a great majority of theregion’s Homo erectus fossils, archeo-logical sites, and chronological inter-pretations. It is worth summarizingthe evolution of chronological frame-works at Sangiran.

A micro-paleontological Plio-Pleis-tocene (current early Pleistocene)chronology was established for San-giran well before WWII.228–230 Afterthe war, a reinterpretation of the evi-dence suggested that the fossilsbelonged to, in current terms, thelate Early-early Middle Pleistocene(�1.0 to 0.4 Ma).84,231–235 However,new microfossil analyses done in the1970s strengthened the initial ‘long

chronology.’ 236–238 Alternatively, thefirst widespread radiometric analy-ses, based on counting fission tracks,yielded young and highly scattereddates.21,239–241 Since the 1980s, agrowing number of 40Ar/39Ar datesand continually refined samplingschemes have generally shown San-giran’s hominin-bearing sequence tohave begun more than 1.5 Ma, possi-bly as much as 1.86 Ma.19,20 Similarchronological debates have beenseen for Arc locales, including Flores,Luzon, and Sulawesi.

The long chronology suggeststhat early hominins arrived inSunda (Sangiran) in the form ofHomo erectus before 1.5 Ma. Anearly form arrived east of Wallace’sLine, to Flores, about 1.1 Ma. Inthe Late Pleistocene, the probableultimate survivors of these earlydispersals are found on Java(Ngandong Homo erectus, �125ka),141 Flores (Homo floresiensis,

�60 ka),184 and Luzon (Homo spindet, �67 ka).156

Marine Isotope StagesThe global MIS provide a cali-

brated environmental frameworkrelating to basic ISEA habitatchange. The MIS system uses oxygenisotope levels to indicate alternatingwarm and cool periods during theQuaternary Ice Age or Pleistocene.Isotope data are drawn from pollenand foraminifera remains in sea-bottom sediments. Deep-sea coresediments also preserve paleomag-netic reversals that can aid in estab-lishing chronology. ISEA deep-seacores also contain air-fall tephrafrom volcanic eruptions and micro-tektites from the Australasian TektiteStrewnfield (Box 5). When subjectedto radiometric age analysis, thesehorizons can be used to refineregional chronology.


larger-bodied stegodons could havedispersed to formerly small land-masses where dwarf species couldhave been overwhelmed.

Mata Menge, Boa Lesa, and Koba-tuwa. These three stratified sites liein the sandy interval above Wolo Segeand Tangi Talo (Fig. 3B).47 MataMenge has a basal date of �880 ka.Boa Lesa and Kobatuwa lie up-sectionfrom Wolo Sege and Tangi Talo. Aseries of tuffs runs through the entiresequence. Fission track dates for cap-ping sediments place them at �700 ka.Among the three sites, major faunalelements include large-bodied stego-don (Stegodon florensis), Komododragon (Varanus komodoensis), giantrat (Hooijeromys nusatenggara), andgiant tortoise (Geochelone). The largestegodon on Flores is somewhatsmaller than the large form on Java.77

With a total of 507 stone tools,Mata Menge has the largest earlyhominin archeological assemblage inISEA.78 Lithic resources werebrought to the site in the form offlake blanks. The more complex toolsshow a reduction sequence based oncentripetal or radial removals on the

blanks (Box 3). The Acheulean-like

elements of Wolo Sege are not pres-

ent at Mata Menge.The Soa Basin Ola Bula formation,

from Wolo Sege to Mata Menge, has

ISEA’s most complete archeological

record. Wolo Sege indicates hominin

arrival at more than 1.0 Ma. The

change in fauna between Wolo Sege

and Mata Menge demonstrates an

EPR-related faunal turnover at about

900 ka. After this turnover, the

Soa Basin shows phylogenetic conti-

nuity in large-bodied animals until

the arrival of Homo sapiens at

12 ka.47

Southern Wallacea Outer Arc

(Indonesia)

South of Flores, the Savu Basin

subduction trench ranges to more

than 3,000 m in depth (Fig. 8).

Sumba and Timor lie on the basin’s

southern arc. At the lowest Pleisto-

cene sea-level lowstands, overwater

channels of 20–40 km separated

these islands from Flores. Both

islands have sites at which Paleo-

lithic stone artifacts and Pleistocene

vertebrate fossils are found in physi-

cal proximity. Nevertheless, age andrelationship remain unclear.

Talau basin (Timor). Pleistocenefluvial conglomerates are exposedalong the Talau River borderbetween Indonesian Timor and EastTimor. During the 1950s and 1960s,localities east and south of Atambua(Indonesian Timor) produced in situfossils and surface stone tools.79,82,84

Taxa included Stegodon timorensis(pygmy stegodon), Geochelone (gianttortoise), and Varanus komodoensis(Komodo dragon).80,81,83,86 Duringthe 1990s, fauna and artifacts wereassociated at two more sites, butaccounts remained preliminary.87 In2015, at Raebia, east of Atambua,the Iowa-Bandung project locatedstegodon, giant tortoise and possiblestone artifacts in relation to bracket-ing tuffaceous lenses. Nearby, atSadi laun, stone artifact clusterswere excavated from deflated terracedeposits.

Noelbaki (Timor). At Timor Island’swest end, near Kupang, Paleolithicstone artifacts appear in coarse flu-vial gravels cut through by modern

Box 5. Pleistocene Catastrophes

Three region-wide catastrophes arevisible in the ISEA early homininarcheological record. The Australa-sian Tektite Impact Event producedthe widespread Australasian TektiteStrewnfield. Cambodia’s Tonl�e Sap,the largest freshwater lake in South-east Asia, may represent the impactcrater (Fig. 1).242 Age for the Austra-lasian Impact is estimated at 0.8 Ma,based on 40Ar/39Ar and fission-trackanalysis of individual tektites. 243–247

Corroborating evidence comes fromthe position of microtektites in deep-sea cores, always below indicationsfor the Brunhes-Matuyama geomag-netic reversal of 0.78 Ma.248–250 Aus-tralasian tektites are known inassociation with early hominin arti-facts and mammalian fauna in vari-ous contexts, including the BoseBasin, Guangxi, China, and the Caga-yan Basin, Luzon, Philippines. In

Bose, the Australasian event mayhave caused widespread, devastatingforest fires which early homininsadapted to through the use of bifacialstone tools.247 The association ofstone tools and tektites in Bose is,nevertheless, in debate.251,252

The Toba caldera of northernSumatra is earth’s largestPleistocene-age volcanic complex.Two of numerous eruptions pro-duced the Oldest Toba Tuff (OTT)and the Youngest Toba Tuff (YTT).New research suggests that OTT rep-resents a total tephra volume of 2,300cubic km,253 which is comparable tothe better-known YTT total volume of2,800 cubic km.149 The AustralasianStrewnfield and OTT are closelyrelated in time. At ODP site 758,northwest of Toba, Australasianmicrotektites and OTT are associatedin ash layers D and E.254 YTT is

found in undersea deposits across theIndian Ocean, as well as the Arabianand South China Sea.255 YTT dates toabout 73,500 years 6 3,000 years148

or 73,000 years 6 4,000 years ago.149

YTT has been implicated in initiatingthe last glacial cycle due to its coinci-dence with not only ice buildup dur-ing this time,256 but populationbottlenecks of Neanderthals in west-ern Eurasia and Homo sapiens inAfrica.257,258 However, the argumentsare weak. YTT traces have not beenidentified at ISEA fossil homininsites. The Middle Solo sites and LiangBua appear to predate the event andthe Callao stratigraphy apparentlypostdates YTT. Given the presentstate of knowledge, it seems thatacross ISEA the YTT event did notgenerate significant bottlenecks orextinctions among mammalianpopulations.259–261


watercourses. In 1978, Soejono sur-veyed at Noelbaki and Noeltarus,reporting choppers, chopping tools,flakes, blades, and proto-hand-axesmade from large flakes.88 In 2015, atthe same localities, the Iowa-Bandungproject differentiated a laminar indus-try with limited patina from largeflake tools with heavy patina. Contex-tual materials are in analysis.

Watambuka and Lewapaku(Sumba). On Sumba’s northeastshore, Watumbaku lies in current estu-arine contexts. In 1978, a stegodon leftmandible (Stegodon sumbaensis) wasrecovered without context.89,90 In2015, the Iowa-Bandung project foundtwo retouched chert flakes in deposi-tional contexts indicating lower(Pleistocene) sea level. In 2012, atLewapaku, 30 km inland, the Wollongong-Bandung Geological Museum projectfound, in situ, tusk fragments, a tooth(Varanus komodoensis), a giantmurine rodent (?Hooijeromys), and abird long bone. A loose comparisonwas made with the fauna of TangiTalo.91

Luzon (Philippines)

The Philippine archipelago hasearly hominin evidence northern-most on the Arc and farthest fromSunda. Deep sea channels currentlyseparate the archipelago from otherISEA landmass systems, suggestingthe presence of dispersal barriersthroughout the Early and MiddlePleistocene.

Paleoanthropological research hasa long history on Luzon. During the1950s, Paleolithic flake tools andextinct fauna were found together inthe Cagayan Basin, and probable dis-persal routes were hypothesized.92

Australasian tektites were also dis-covered at several locations.93 Duringthe 1970s, researchers at theNational Museum of the Philippinessurveyed many more Cagayan co-occurrences of stegodon and arti-facts.94 More recently, fossils of sev-eral mammal families have beenfound in archeological contexts,including small stegodon (S. luzonen-sis), and the tektite association hasbeen confirmed.95,96 Recent researchalso suggests that the Luzon Middle

Pleistocene flake tool assemblages fitwithin the Large Flake Acheulean.97

Cagayan Basin. At Luzon’s northernend, the Cagayan Basin is a 250 x 80km subduction zone feature with 10vertical km of sedimentary infill (Fig.3C). The upward-coarsening volcani-clastics reflect the tectonic and vol-canic evolution of the adjacentCordillera Central volcanic arc. Theupper 900 m comprises two forma-tions of interbedded fluvial and pyro-clastic deposits, the Ilagan (lower500 m) and the Awidon Mesa (upper400 m). The sequence consists offour depositional environments:meandering stream, braided stream,lahar and pyroclastic flow, and ash-fall deposits.98 Cagayan has scores oflocalities with stone tools and, insome cases, tektites and Pleistocenefauna.99 Field work has recentlybeen resumed in the Cagayan Basinwhere the contemporaneous occur-rence of stone tools, tektites, andPleistocene faunas has beenconfirmed.100

Enrile Southern Quarry. An impor-tant Awidon Mesa formation localityis Southern Enrile Quarry, nearPe~nablanca. Here, a Danish-Australian team is bracketing thefossil and artifact-bearing level using40Ar/39Ar on volcanic elements andluminescence on low-temperaturesediments. Preliminary 40Ar/39Arages range from the late Early andMiddle Pleistocene to 0.4 Ma as thesecure youngest age.101 It is reasona-ble to entertain an Early PleistoceneRevolution (EPR) time frame for theartifact-fauna-tektite association.

Arubo Basin. Approximately 300km south of Cagayan, Arubo is acomplex of open sites in the SierraMadre foothills of Central Luzon.102

A morphologically heterogeneouslithic assemblage has been collectedhere, primarily from sites out of geo-logical context.97 The site complexlies close to a chert boulder deposit,which served as the raw materialsource. The stone tool assemblagesinclude the Large Flake Acheulean(LFA) hallmarks of bifaces, cleavers,flake cores, retouched flakes, andchoppers (Fig. 5). Microscopic use-

wear analyses suggest curation andreuse.103,104 The Arubo complex ageremains unknown, but is likely to becorrelative with the LFA features atthe Cagayan sites.

The Stegodon-artifact-tektite (Box5) association at Awidon Mesa For-mation is crucial for understandingLuzon early hominin arrival (Fig. 5).This association suggests that earlyhominins arrived at Luzon by 800ka, probably bringing LFA technol-ogy. It remains to be seen whetheror not the Luzon dispersal relates tothe Soa Basin Middle Pleistocenefaunal turnover. It must be notedthat in the Cagayan and AruboBasins, the Middle Pleistocene stonetechnology is not easily differenti-ated from the Tabonian industries ofthe late Upper Pleistocene found onPalawan Island, which are associatedwith Homo sapiens.94

MID-BRUNHES EVENT OR LATEMIDDLE AND LATE PLEISTOCENE

MIS 12/11 (�480-360 ka) began anew pattern within the 100-kyr cli-mate cycle. Glacial phases becamelong (70-90 ka) and very cold, whileinterglacials became short (10-30 ka)and warmer.105 This transition, theMid-Brunhes Event (MBE), beganthe four large-amplitude glacial-interglacial cycles that have struc-tured global climate to the pres-ent.106 With MBE, earth’s climatebecame more orderly, predictable,and extreme.107

MBE had significant global conse-quences. MIS 12 exhibited severe cool-ing effects.108,109 MIS 11 marked thelongest, warmest interglacial, with sealevel rising to 20 m above the presenthighstand.110 The MBE 12/11 complexprobably aided in both dispersing andisolating ISEA large mammals. It isworth noting that in western Eurasia,the appearance of Neandertal traits iscorrelated with the onset of MIS 11.111

Extreme interglacial sea level high-stands also came during MIS 9 and 5e.

On mainland China, in conjunc-tion with MBE, the Hulu andDangge cave flowstones record thebeginning of monsoon rainfall. Theflowstones record the dry monsoonsdirectly associated with the com-mencement of massive periodic


North Atlantic glacier ice calvingknown as Heinrich events.112,113

These events produced cold wintersin Europe and dry monsoons inChina.114,115 Another consequence ofHeinrich events was the southwardshift of the Tropical Rain Belt.116

The sequence of events, stretchinghalfway around the globe, indicatesthe swiftness of glacial climatechange.

The Mid-Brunhes Event markedyet another significant change inregional erosion and sedimentationpatterns. Some Miocene limestonemassifs developed karst landscapes.Central Java’s Southern MountainsMiocene coral beds (Gunung Sewu)provide an example. At Punung III inGunung Sewu, a U-series age of 4926 38 ka for the lower flowstone pro-vides a minimum age for uplift.117

This determination corroboratesother Middle Pleistocene uplift esti-mates relating karst development inthe Southern Mountains.118–120

Early hominins found shelter inthese newly opened caves. The SongTerus cave fluvial stage documentsthe accumulation of stone artifactsbefore 300 ka. Liang Bua opened forinfilling about 195 ka and shows evi-dence of hominin habitation shortlythereafter. The evolution of Callaocave is not yet clear.

The Mid-Late Pleistocene bound-ary (MIS 5e, 130-120 ka) was a glob-ally important climatic event,represented in Paleolithic sitesacross northwestern Eurasia. How-ever, sedimentary sequences inSoutheast Asia do not seem to regis-ter the Last Interglacial in significantways. One Late Pleistocene catastro-phe, the Youngest Toba Eruption, isnot identifiable in sedimentarysequences throughout ISEA (Box 5).Table 3 shows the events, sites,lithics, and fauna present during theMid-Brunhes Event, late MiddlePleistocene, and Late Pleistocene.

Sunda

Trinil, Java

Recent observations from Dubois’historical faunal collection at theNetherlands Museum in Leiden46

suggest that Homo erectus may have

TAB

LE3.

Mid

-Bru

nh

es

Eve

nt

Syn

op

ticC

ha

rt

MIS

sta

rt

�k

a

en

d

�k

ae

ve

nts

ag

e

ka

Java

&[

Ma

lay

sia

]Fl

ore

s

tre

nd

ssi

tes

lith

ics

Ste

go

do

nsi

tes

lith

ics

Ste

go

do

n

112

Ho

loc

en

e2

28

12

Flo

res

eru

ptio

ns

17

362

28

474

62

Yo

un

ge

stTo

ba

73

[Le

ng

go

ng

]fla

kes

5120

74

Ng

an

do

ng

flake

sS.

trig

on

oc

ep

ha

lus

Lia

ng

Bu

ara

dia

lc

ore

sS.

flore

nsi

sS.

insu

laris

MIS

5-2

:IT

CZ

pla

ce

me

nt

for

SEM

on

soo

n.

Dry

co

nd

itio

ns

thro

ug

ho

ut

ISEA

5e

130

120

Last

Inte

rgla

cia

l;h

igh

sea

leve

ls6

190

130

Lia

ng

Bu

ase

dim

en

tatio

n7

250

190

8300

250

Son

gTe

rus

flake

s(P

ac

ita

nia

n?

)S.

trig

on

oc

ep

ha

lus

dry

ing

inN

ew

Gu

ine

a&

Au

stra

lia9

340

300

Son

gTe

rus

sed

ime

nta

tio

n10

360

340

hig

hin

terg

lac

ialse

ale

ve

ls11

420

360

MB

E420

seve

rec

oo

ling

100-k

yr

ob

liqu

ity

cyc

le:

co

ldg

lac

ials

(70-9

0ka

);w

arm

inte

rgla

cia

ls(1

0-3

0ka

)


occupied Trinil during the MiddlePleistocene, displaying advanced cog-nition and cultural behavior. Withinthe collection, the Pseudodon shellassemblage represents adult large-sized specimens (80-120 mm inlength) from varied riverine settings.This suggests that individual shell-fish were selected for consumptionbased on large size. One shell valveshows signs of modification byretouch, possibly resulting from useas a cutting or scraping tool. Anothershell displays a zigzag pattern ofgrooves on the central part of theleft valve. The marks are attributedto the engraving of meaningfulabstract patterns on an attractivesubstrate.121 In other words, Trinilhominins are thought to have inten-tionally marked the shell.

Luminescence and 40Ar/39Ar analy-sis on the shells’ adhering sedimentgive a maximum of 0.55 Ma and aminimum of 0.44 Ma,46 or abouthalf the age estimate based on analy-sis of the associated mammalianfauna.44 This new age would seem topull Trinil away from the Sangiranfossils and an early Middle Pleisto-cene Ngebung cultural affiliation.Should the shell marking prove fac-tual, does it indicate cognitive abil-ities expectable in the later Java sitesin the Song Terus karst and the Mid-dle Solo Terraces? On date alone, itis tempting to draw a comparisonwith mid-Middle Pleistocene earlyhominin dentognathic fossils fromHexian, eastern China, and Penghu,Taiwan. Trinil at a mid-Middle Pleis-tocene age may represent a time inwhich Homo erectus diverged in sev-eral regional contexts.

The Dubois collection zigzag shellcould represent part of a regionaltrend toward early hominin culturaldiversification. Nevertheless, the Tri-nil shell marking has a complexitynormally associated with early LatePleistocene Homo sapiens inAfrica.122,123 It is possible that diver-sifying Middle Pleistocene Homoerectus had similar cognitive ability,but the find is unique. The problemremains that the Trinil Hauptkno-chenschicht is a complex deposit,apparently a mixed sequence offlood-borne overbank accumulations.Bones, shells, and sediments of var-

ied chronology may have merged in

this context.121 Given the poorly

known nature of the site and the

primitive morphology of the Trinil

calotte, we hesitate to grant

advanced symbolic behavior to the

group represented by the calotte.

Gunung Sewu, Java

The Southern Mountains of East

Java lie on the Indian Ocean coast,

100-150 km south of the Solo Basin.

The area has been of geological

interest since the 1830s and of arche-

ological interest since the 1930s.124

In 1935, von Koenigswald and Twee-

die collected a range of flake and

pebble tools from the Baksoka riv-

erbed, later described as the Pacita-

nian industry.125 Since the 1990s,

Gunung Sewu research has focused

on cave and rock shelter habitations

and the use of local lithic resources,

including chert, jasper, limestone,

and meta-limestone.124 Gunung

Sewu stone sources may have pro-

vided the raw materials for Sangiran

sites such as Ngebung.

Song Terus. Most Gunung Sewu

caves have yielded Late Pleistocene

and Holocene deposits, but one,

Song Terus, features an important

Middle Pleistocene infill. In 1953,

Soejono and Basuki excavated fauna

and an archeological industry involv-

ing stone, bone, and shell.52 In the

late 1990s, an Indonesian-French

team excavated more than 15 m of

stratigraphy in two units.126 The

lower unit consists of flood alluvium

from the 12-m terrace. The layers

contain fauna (rhinoceros, tapir, and

cervid) and flake tools.85,126 Com-

bined U/Th-ESR analysis indicates

that the lowest archeological levels

arrived �300 ka.127 The flake tools

may relate to classical Pacitanian

lithics (Fig. 6).126 The fluvially

derived archeological deposits

increase in density after 180 ka. By

80 ka, the local stream at Song Terus

had entrenched below the cave

entrance and a more typical cave

infilling took over. Archeological

remains include hearths and fauna,

but few lithics. This seems to repre-

sent sparse inhabitation and may be

indicative of waning Sunda Homoerectus populations.126

Other Gunung Sewu caves, suchas Song Klepek and Braholo, aid inunderstanding late Homo erectuslithic technology and landscape use,and may eventually shed light on theterrace lithic industries such as thePacitanian and Sangiran FlakeIndustry. The deep Gunung Sewucaves may be especially importantfor understanding hominin adapta-tions during MIS 6 (150-130 ka),when the Southeast Asian mainlandwas exceptionally cold and dry. Atthis time, the equatorial insularprovince may have served as an earlyhominin refuge.

Middle Solo Terraces, Java

As the Solo River cut through theKendeng Hills, it entrenched withinPleistocene terraces. In 1931, Geologi-cal Survey of the Netherlands geolo-gist Carel ter Haar discovered a bonebed near the base of the Solo River

Figure 6. Two bifaces from the Baksoko val-ley, Gunung Sewu, Java (modified afterG.-H. Bartstra).125


20-m terrace. Ngandong sat on thecutbank of an acute river bend withinthe Kendeng Hills. From 1931 to1933, the Survey excavated the bonebed within a volcaniclastic sand stra-tum.128–130 Ngandong produced 12“Solo Man” cranial remains.

Since the early 1970s, fourNgandong-like calvaria have beenfound 15–20 km upstream fromNgandong, a few kilometersupstream from the Trinil locality.The calvaria are chance finds in theeroding Solo River bank; the sedi-mentary details of their provenanceare as yet unknown. As a series, thefossils are generally called by thenames of their administrative munic-ipalities, Sambungmacan–Ngawi,

and are more specifically identifiedby the three villages nearest theirfind spots (Table 4).

Middle Solo cranial morphol-ogy. Compared to Sunda EarlyPleistocene Homo erectus, the Sam-bungmacan-Ngawi specimens showderived features, including a relativelyhorizontal supraorbital torus thatthickens laterally, continuous supra-meatal/supramastoid crests, and awell-defined occipital torus and supe-rior nuchal line.131 Cranial walls arerelatively tall and the coronal profile is“roundedly tent-shaped.”131 Cranialcapacity is increased.131–133 The seriesexhibits minor postmortem distortionsor surface erosion.

The Ngandong hominins rangefrom cranial vault fragments to anearly complete calvaria retainingdelicate ethmoid and sphenoid struc-tures.134 The Ngandong calvaria arelarger and more robustly built thanthe Ngawi-Sambungmacan series(Fig. 7). The lateral walls of thebraincase are more vertically ori-ented, helping cranial capacity rangeupward to 1,250 cc. Raised temporallines on the cranial vault give a puffyappearance to the vault when viewedfrom the front. The strongly devel-oped supraorbital torus does notform a continuous bar, but rathermeets in the glabellar region in a dis-tinct depression. The pronouncedoccipital torus has greater rearwardprojection.131

Ngandong taphonomy. The 1930sexcavations produced �25,000 verte-brate fossils, including at least 10terrestrial mammals: buffalo (Buba-lus), cattle (Bibos), deer (Cervuspalaeojavanicus), hippopotamus(Hexaprotodon), leopard (Pantherapardus), muntjac (Muntiacus), pig(Sus terhaari), rhinoceros (R. sondai-cus), stegodon (S. trigonocephalus),and tiger (Panthera tigris). Bovidsrepresent more than half of theassemblage.135 Broken and disarticu-lated elements greatly outnumberedwell-preserved specimens and par-tially articulated skeletons. No cutmarks have been reported and fewverifiable artifacts appeared inexcavation.

Recent excavations have isolatedthe extent of the 1930s excavationand located the bone bed stratumwithin. The Homo erectus specimenscan now be placed surely within theoriginal facies C of the basal fossilif-erous horizon.130,134,136 Bone bedsedimentary dynamics are also underinvestigation. The deposit, compris-ing poorly sorted, high-energy fluvialsand and gravel, shows hyper-

TABLE 4. Sambungmacan-Ngawi Calvaria Sites

Village Location Date ID Anatomy

Poloyo – 1973 Sm 1 calvariaPoloyo – 1973 Sm 2 tibia fragmentMlale-Cemeng 4 km upstream from Poloyo 1977 Sm 3 calvariaMlale-Cemeng 4 km upstream from Poloyo 2001 Sm 4 calvariaSelopuro 6 km downstream from Trinil 1987 Ngawi 1 calvaria

Figure 7. Homo erectus calvaria (Solo Skull V, Ng 6): Ngandong, Middle Solo Valley,Java, Cranial capacity: 1,251 cc.275 Courtesy of the American Museum of Natural His-tory, New York.


concentrated flow features typical ofa lahar event. The volcaniclasticsource appears to be an andesiticcone located about 50 km away.134

With this new information, ataphonomic hypothesis can besketched, its events includingupstream multi-species aggregation,mass death and relatively speedy car-cass decomposition, and mass flowof osseous elements downstream toconcentration and burial at Ngan-dong. The aggregation may havebeen caused by drought or volcaniceruption; mass death may be relatedto localized ash fall. During lahartransport, the carcasses were disar-ticulated and individual bones bro-ken, but surfaces and edges were noteroded.134

Middle Solo Terraces Ages. The ageof the 20 m terrace sediment constit-uents is yet to be resolved. The his-torical sequence of age analyses ispresented in Table 5.137–141

The oldest dates push these fossilsback into the Middle Pleistocene.The youngest results (53 ka – 23ka)138 straddle the 47 ka arrivalthreshold for Homo sapiens to ISEAand Sahul. The youngest dates sug-gest that Homo erectus survived thearrival of Homo sapiens for a signifi-cant time. It is reasonable, neverthe-less, to reject the earliest and latestages as anomalous. The bulk of theages suggest that the hominins atMiddle Solo terraces generally relateto the Last Interglacial (MIS 5e), fall-ing just before, during, or just after.The general consensus is that theMiddle Solo terraces represent thelatest form of ISEA Homo erectus.140

The question is open as to whetheror not arriving Homo sapiensencountered Homo erectus of theNgandong type.142

Lenggong Valley, Perak

(Peninsular Malaysia)

Since the 1930s, several LenggongValley localities have produced sim-ple stone tool assemblages collec-tively known as the Tampanian.143

As surface finds, the Tampanian wasthought to be an early hominin tech-nology similar to the Pacitanian.144

The site of Kota Tampan is one ofseveral localities that show a tuffa-ceous stratigraphy without faunalpreservation. Evidence is now con-clusive that the site lies within theYoungest Toba Tuff (YTT).145–147

YTT dates to about 73,500 years 6

3,000 years148 or 73,000 years 6

4,000 years ago (Box 5).149 Otherexcavated Lenggong sites are strati-graphically younger. One muchyounger site includes a Homo sapi-ens skeleton.150

The Lenggong Valley sites and theTampanian are difficult to place inhuman biogeographic context. TheYTT date puts Kota Tampan withinthe realm of possibility as an earlyhominin site; however, the stonetechnology associated with the sitecould represent either early homi-nins or modern humans. The YTTdate has been used as evidence of apre-YTT eastward dispersal of Homosapiens.151

Eastern Island Arc

Wae Racang Karst, Flores

On northwestern Flores, the Man-garri limestone massif (Miocene age)extends more than 500 m above sealevel (Fig. 3B). The local karst systembegan developing about 600 ka asthe Wae Racang River incised morethan 100 m into the massif. Fiveriver terraces record this evolution.The Liang Bua cave site is the result

of the river exposing, then invadingthe karst system. The river now lies30 m below and 200 m distant fromthe cave.5

Liang Bua. The cave is 14 km northof Ruteng and 25 km from the northcoast. Stone artifacts began accumu-lating at about 190 ka. After 100 ka,channel erosion created relief withinthe soft debris. Remnant areas ofhigher ground later became a focusfor hominin habitation from 74–61ka.152 There is no direct indicationof YTT in the cave sediments. TheLiang Bua archeofauna sequenceranges from �95 ka to the pres-ent.153,155 Recently, a depositionalhiatus has been identified between�60 and �17 ka.153

The Liang Bua archeofauna con-sists of unfossilized but well-preserved mammal, bird, reptile, andmollusk remains. The pre- and post-hiatus assemblages are distinct inspecies representation. For the earlysequence, that associated with Homofloresiensis, large vertebrates includegiant tortoise (Geochelone), threegiant rat species (Papagomys, Spelae-omys, or Hooijeromis), and Komododragon (Varanus komodoensis). Asimilar faunal association is seen atthe Soa Basin sites about 700,000years earlier. However, Liang Buauniquely preserves a very small steg-odon (Stegodon florensis insularis),thought to be the dwarfed descend-ant of the large-bodied EPR-arrival,Stegodon florensis, as well as adiminutive hominin (Homo floresien-sis).153 The origin and unique mor-phology of Homo floresiensis hasbeen the subject of numerous inter-pretations (Box 6).

Excluding the hominin fossils, theLiang Bua fauna is characterized byphylogenetic continuity and low

TABLE 5. Middle Solo Terraces Historical Ages Analyses

Year Methodology Age

1939 Von Koenigswald attributes the Ngandong 1930s excavation fauna Upper Pleistocene137

1988 U-series on Ngandong bone fragments �165 ka129

1996 ESR/U-series on bovid teeth from the Ngandong 1930s excavation area 53-27 ka138

2007 luminescence and U-series on Punung fauna breccias* 128 6 15 to 118 6 3 ka117

2008 gamma-ray spectrography on Ngandong/Sambungmacan hominin fossils 70-40 ka139

2011 40Ar/39Ar incremental heating & ESR/U-series on Ngandong and Jigar fauna 546.6 6 12 ka140

ESR/U-series min age: 143120/-17 ka140

2014 red thermoluminescence (red TL) on Ngandong bone bed fluvial sediments �130-102 ka141


species richness, as well as by rela-tively large and small body sizes.These are features of “impoverishedand disharmonic insular faunas” typ-ical of isolated oceanic islands.6

Within this context, the Homo flore-siensis skeleton exhibits small stat-ure, a small brain, relatively longarms, robust lower limbs, and longfeet. In sum, Homo floresiensis skele-tal features resemble those of small-bodied taxa on Flores and across theArc. Small bodies helped these formsremain successful as long-term mem-bers of insular Arc faunas (Box 2).6,7

The early Liang Bua stone technol-ogy shows a radial or centripetalreduction sequence similar to thatseen at Mata Menge during the Mid-dle Pleistocene.154 The stone toolsare most closely associated with thebutchered stegodon remains, primar-ily dental and skeletal elements ofjuveniles.155 Continuity in the stoneartifact technology, Mata Menge toLiang Bua (800 kyr), strengthens theimage of Flores as an isolated earlyhominin territory.78,154 Archaeo-faunal continuity seems to corrobo-rate the trend.

Homo sapiens arrived in ISEA,

likely including Flores, by 47 ka. At

Liang Bua, nevertheless, modern

humans appear after deposition

resumes at �12 ka. The Homo sapi-ens archeofauna sequence has fresh-

water mollusks, including Thiaridae

(Thiara granifera and Melanoidestuberculata) and Neritidae (Neritinapulligera, Neritodryas cornea, Nerito-dryas dubia, Septaria porcellana, and

Clithon squarrosus).155 Modern

humans also introduced a range of

exotic animals to the island, includ-

ing the Sulawesi warty pig (Sus cele-bensis) and the Eurasian pig (Susscrofa), long-tailed macaque (Macacafascicularis), Javanese porcupine

(Hystrix javanica), and masked palm

civet (Paguma larvata).155 Only the

Komodo dragon (Varanus komodoen-sis), an opportunistic predator and

scavenger, is represented throughout

the archeofauna sequence.

Cagayan Basin, Luzon

Callao Cave. The Cagayan’s eastern

flank has a significant Miocene karst

massif (Fig. 3C). In 2003, Armand

Mijares began excavating at theCallao Cave opening.156 In 2007, aPhilippine-Australian partnershipexpanded the work. The excavationencountered a rich fauna includingnative brown deer (Cervus mariannus;90% of the identifiable bone frag-ments), Philippine warty pig (Susphilippensis), and an unspecifiedextinct bovid. Element representationand fragmentation for the cervidsindicated that both whole and partialcarcasses were brought into the cavefor further processing. Some bonesshow cut marks, but no stone toolswere found.8

Near the base of the excavated area,Layer 14 yielded a carbonized brecciawith relatively dense fauna. Amongfragmentary hominin limb bones, thebreccias contained a complete thirdmetatarsal. Two cervid teeth fromLayer 14 were dated. U-series abla-tion on one tooth produced a mini-mum age estimate of 66.7 ka. ESR onthe other tooth corroborated theresult.8 In 2014, further excavation inLayer 14 produced more hand andfoot bones and a series of upper teeth.The dentition is of archaic character.9

Box 6. Homo floresiensis Skeleton

The Liang Bua skeleton (LB1),the smallest known fossil hominin,expresses a mosaic of primitive andderived features that are absent inother early hominin groups.262

Fragmentary remains have beenrecovered from at least nine individ-uals.263,264 Our description of Homofloresiensis is based on partial skele-ton LB1 and mandible LB6. LB1exhibits a marked reductive trend inits facial skeleton, with extremelysmall overall cranial size, a primi-tive low and anteriorly narrow vaultshape with thin cranial bones, a rel-atively prognathic face, and smallteeth.184 Nevertheless, the mandi-bles of Homo floresiensis (LB1 andLB6) are buttressed, as seen in theholotype of Homo habilis (OH7).265

The postcranial skeleton exhibitssome primitive features, includingflared hipbones; short collarbone;forwardly positioned shoulder

joint263; shortened femur and tibia;trapezoid, scaphoid, and capitatewrist bones that are primitive andresemble those of Homo habilis(OH8)266; and a foot exhibitingprimitive features including longlateral toes, a short hallux, and theabsence of a well-defined mediallongitudinal arch, resulting in flatfeet.267 The only relevant wrist ele-ment known for Homo erectus ineither Africa or Asia is one damagedlunate from Zhoukoudian.268 It istherefore impossible to compare theLB1 wrist morphology with that ofHomo erectus. It is distinctly possi-ble that the wrist morphology ofearly Homo erectus resembled thatof Homo habilis (OH8), and thuscould have been a likely precursorto Homo floresiensis. This is moreplausible than a proposed transcon-tinental migration of Homo habilis.

A decade after discovery, the evo-lutionary processes responsible forthe small size and unique morphol-ogy of the LB specimens are stilldebated.265 An extreme interpreta-tion relates LB1 to a pathologicallydwarfed Homo sapiens individual.Jacob and colleagues suggest thatLB1 represents a “pygmoid austral-omelanesian” with developmentalabnormalities.179 Other hypothe-sized conditions include Laron syn-drome,269 cretinism,270 and Downsyndrome.271 However, LB1 mor-phology does not definitively reflectany known systemic pathology. Cur-rently, the LB1 stratigraphic contextis being reevaluated.50 A probableolder last occurrence for Homo flor-esiensis at Liang Bua, in excess of60 ka, would preclude the patholog-ical Homo sapiens hypothesis.


The initial interpretation of theCallao metatarsal went in two direc-tions. The small dimensions and grac-ile morphology were linked to small-bodied Homo sapiens, including mod-ern Philippine Negrito populations.Alternatively, the fossil drew compari-son with small early hominin speciessuch as Homo habilis and Homo flore-siensis. The new Callao finds corrobo-rate the early hominin comparisonand strengthen the hypothesis for asecond colonization east of Wallace’sLine. It is tempting to link the originsof these finds with the EPR-age Caga-yan Basin dispersal involving homi-nins, large-bodied stegodons, giantrats, and other insular species. TheAustralasian Tektite Impact age sug-gests that such an event would dateto at least 800 ka.

Maros Karst, South Sulawesi

Some 80 km southwest of theWalanae sedimentary basin lies theMaros limestone karst system. Here,local stream courses follow intersect-ing joints to form plateau-like hillmasses. In areas of maximum pla-teau dissection, steep-sided towersare isolated on alluvial plains.157 Thetowers have many caves and rockshelters, some of which have prehis-toric archeology.

Recent excavations in several caveshave exposed cultural levels extend-ing back more than 35 ka. Stonetechnology is based on small bipolarcores that can resemble very smallbifaces. Archeofaunas have smallmammals, including monkeys(Macaca sp.), bear cuscus (Ailuropssp.), and Celebes warty pigs (Sus cel-ebensis). Freshwater gastropods(Tylomelania perfecta), fish, and birdsare also represented.158,159 Pigment-based rock art is now dated to 39.9ka.160

At one Maros cave, an upperarcheological complex ranges backto �41 ka. Well below, a lowerassemblage is quite different. Thestone technology is based on mini-mally reduced cores and retouchedcobbles yielding flakes and debitage.The fauna has Elephas and Stegodonand no smaller species. The levelscontain no pigments. The depositsare older than �53 kyr ago and may

be significantly older. This is a majorchange in the archeological column.The most economical interpretationis that at �41 ka H. sapiens replacedan earlier population of archaichominins.161 The cave environmentsof the Maros karst may represent thebest preservation of levels indicatingthe arrival of H. sapiens and theextinction of an as-yet undeterminedearly hominin population.

DISCUSSION

With basic space and time data inhand, we can turn to basic issues ofISEA early hominin biogeography.Two assumptions guide us. First,early hominins arrived at ISEA asgeneralized hunter-scavengers inte-grated within mainland Eurasianmammalian faunas.162 As such, hom-inins directly responded to the move-ments of nonhuman fauna members.Second, within insular contexts,mammalian faunas were particularlysensitive to environmental processes,including dispersal, isolation, vicar-iance, relict survivorship, and geneticdrift. During a million-and-a-halfyears of shifting ISEA habitats, earlyhominins responded to many insularopportunities and constraints. Here,we address current issues relevant tothe dispersal and isolation of earlyhominins and their extinction inrelation to the arrival of Homosapiens.

Dispersal

Overwater Transit. Even at maxi-mum sea-level lowstands, deep-waterchannels separated Sunda and East-ern Island Arc landmasses. Headingeast from Java, the Lombok andKomodo Straits ranged from 20 to35 km at maximum lowstands.163

Farther north, between Borneo andSulawesi, the Makassar Strait wasnever less than 40-km wide. Sulawesiand Flores are separated by a mini-mum of 60 km of sea with cross-currents (Fig. 8). It is unlikely that adry-land connection ever existedacross the Mindoro Strait to connectBorneo with Luzon via Palawan.164

Presently, strong north-to-southflow-through currents assist in iso-lating populations on either side ofWallace’s Line.165,166 In crossing

overwater to the Arc, Eurasian faunacould have arrived from variousSunda sources via current Java orBorneo.32,167

Alfred Wallace was the first to dis-cuss overwater fauna movementswith waif dispersal among the morecommon means.168,169 Waifs areindividuals or small groups sweptaway from one environment andtransported by water or air currentsto a new territory.170 Some terres-trial forms, such as stegodon, gianttortoise, and Komodo dragon couldfloat or swim following the prevail-ing currents. Hominins more likelyrafted, either accidently or purpose-fully.7 Overwater waif dispersal ren-ders unlikely any subsequent geneexchange with the source popula-tion.171 Such long-term isolation onsmall islands is the basis for insularevolution.

Isolation

With the late Early PleistoceneTransition (MIS 23 �900 ka), theglacial-interglacial pulse became lon-ger and more extreme. Early MiddlePleistocene interglacial high sea lev-els may have been the first to isolatehominin groups across ISEA, espe-cially in the Arc. The Solo, Soa, andCagayan basins show major faunalturnovers during MIS 23-20. EPRalso marks the last fossil evidence ofEarly Pleistocene Homo erectus asdefined at Trinil and Sangiran. Withthe Mid-Brunhes Event (MIS 11�450 ka), regional karst systemsbegan to develop. By MIS 9 (�300ka), caves were accumulating earlyhominin living debris and becomingan important record of fossil andarcheological evidence. By the LastInterglacial (MIS 5e �125 ka), ISEAearly hominin groups were appa-rently isolated in at least three areas.Two lines of evidence illuminatesome details of Middle and LatePleistocene isolation.

Stegodontidae. The extinct familyStegodontidae comprised Asian rela-tives of mammoths and modern ele-phants. The fossil diversity ofstegodon species is greatest nearYunnan, in southern China, theirpresumed area of origin. Pliocene


Figure 8. ISEA early hominin sites in bathymetric and geographic context including all geographic and a site names used in this paper.We followed Huxley’s modification of Wallace’s Line208 as illustrated in Cooper and Stringer204. [Color figure can be viewed in theonline issue, which is available at wileyonlinelibrary.com.]



fossils are found across southernChina and Japan. Pleistocene fossils

are known in North China, South-east Asia, and India.172,173 In Pleisto-cene ISEA, Stegodon became adominant faunal element in large-and small-bodied forms.85,174,175

Large-bodied Stegodon trigonoce-phalus is first documented in themiddle Sangiran Formation (c. 1.6Ma) of Java.27 Presumably fromSunda, and with the aid of sea-levellowstands, Stegodon dispersedthroughout ISEA. Earlier Pleistocene

dwarfs appeared on Flores (S. son-daari), Sulawesi (S. sompoensis),Timor (S. timorensis), Sumba (S.sumbaensis), and Mindanau (S.mindanensis).32 Later in the Pleisto-

cene, on Flores, a medium-bodied S.florensis probably gave rise to theLate Pleistocene small-bodied S. flor-ensis insularis. The latter is knownfrom the upper levels of Liang Bua,

where dwarf hominins may havehunted or scavenged this and otherdwarfed forms.155

Dwarfing among ISEA Stegodonti-dae began during the earlier Pleisto-cene and continued throughout the

epoch.32 The great preponderance ofdwarfed forms is found east of Wal-lace’s Line. Here, stegodon sizereduction may be related to theincreasingly high interglacial high-stands commencing with the EPR.

Body size reduction may in partexplain the long-term success ofselect large mammals, includingstegodon and hominins in the Arc.176

West of Wallace’s Line, on Sunda,

there is no evidence of size reductionamong hominins. Between the EPR(MIS 23 �900 ka) and the Last Inter-glacial (MIS 5e �125 ka), the Javahominins, as indicated by cranial

and dental fossils at Sambungma-

can–Ngawi and Trinil (if later), didnot decrease in size, nor is thereindication of endemic dwarfing inSunda stegodon.

There are several associationsof stegodon fossils with early homininartifacts, especially east of Wallace’sLine. Minimally, such co-occurrencessuggest that on the smaller Arc land-masses, early hominins and stegodon-tids frequented similar habitats.Maximally, as may be the case atLiang Bua, the association indicatesthat early hominins scavenged orhunted these proboscideans. Looselyspeaking, the stegodon-artifact co-occurrence sequence appears toreflect early hominin dispersal acrossISEA. The dispersal and isolation ofPleistocene ISEA stegodon may there-fore serve as a heuristic model forinsular development among ISEAearly hominins. Table 6 summarizesthis evidence.

With EPR (MIS 24-22, �1.0-0.9Ma), glacial period sea-level low-stands and aridity became moreextreme. For large mammals, includ-ing hominins and stegodon, ariditymay have prompted dispersal whilesea-level lowstands enabled it. EPRthus provided the first real means toleave Sunda for the Arc and to setup Arc islands up first as glacialperiod refugia, then as interglacialperiod bottlenecks. A similar EPRforcing effect, to provide for earlyhominins and elephants (Elephasand Mammuthus) dispersing in tan-dem, has been suggested for south-ern Europe.60,61,177

Homo floresiensis insularity. Amongthe larger vertebrates, small-bodiedspecies have relatively more successin overwater dispersal and bettersurvival potential on resource-limitedislands.170 With its small, specialized

morphology, the LB1 skeleton seemsto show signs of insular evolutionary

development. We can imagine thatHomo floresiensis is the insular resultof overwater dispersal of either oftwo forms, a small-bodied arrivalthat became further specialized or alarger-bodied arrival that became

dwarfed.At discovery, LB1’s remarkably

small brain size (417 cc) caused someresearchers to argue that LB1 was amicrocephalic modern human178 or apygmoid Australomelanesian modern

human with developmental abnormal-ities179 (Box 6). There is now convinc-ing evidence that LB1 followed aninsular evolutionary path for reducedbrain size demonstrated in other

mammalian lines. Three recent stud-ies bear directly on Homo floresiensisbrain size: island-dwarfed hippos;foxes, mice, and humans; and callitri-chids (marmosets and tamarins).

Weston and Lister180 compared scal-ing models for dwarfed hippos andtheir mainland ancestors. Dwarfedspecies have significantly smallerbrains, in relation to cranial size, than

predicted from scaling mainlandforms. Observing brain size reductionin multiple mammalian lines,Schauber and Falk181 concluded thatthe Homo floresiensis brain could beproportionally dwarfed from a larger-

bodied ancestor having similar rela-tive brain size. Montgomery andMundy182 correlated callitrichid brainsize reduction with a slowdown of theprenatal growth rate. Based on the

callitrichid example, the brain size ofHomo floresiensis may have been sub-ject to selection pressure at earlystages of development.

The LB1 wrist has been comparedwith that of Homo habilis, known

exclusively from East Africa (Box 6).

TABLE 6. Pleistocene ISEA Stegodon-Artifact Co-occrrence Chart

Region Island Site Geological context Technology Taxon Age Reference

Sunda Java Bukuran Sangiran Fm shell S. elephantoides 1.6 Ma 19,32Arca Flores T. Talo1W. Sege Ola Bula Fm flake S. sondaari 1.3 Ma 48Arc Sulawesi Cabenge Walanae Basin flake S. sompoensis Early Pleist 48Arc Luzon numerous Cagayan Basin LFA S. luzonensis 0.8 Ma 96Arc Luzon numerous Arubo Basin LFA S. luzonensis 0.8 Ma 96Arc Timor Atambua Weaiwa Fm flake S. timorensis n/a 79,83Arc Timor Motaoan Weaiwa Fm flake S. timorensis n/a 87

aEastern Island Arc


Late Pliocene or earliest PleistoceneISEA dispersal of a small African hom-inin has thus been suggested.183 Twolines of evidence weaken the Africaconnection. First, LB1 has advancedcraniodental features not seen in anyAfrican early hominin.184,185 Second,there is no fossil evidence across main-land Eurasia for a Homo habilis dis-persal out of Africa.

Alternatively, Homo floresiensisrepresents a diminutive Eurasianarrival. Here, the Dmanisi fossils(Republic of Georgia, southwest Eur-asia) reflect an Early Pleistocene (1.8Ma) paleodeme186 as a potentialISEA donor. Dmanisi Homo erectusfeatures include a small brain (546–780 cc capacity) and small body(145-166 cm height, 40-50 kgweight).3 LB1 cranial capacity (417cc) is not significantly smaller thanthat of Dmanisi skull 5 (D4500) (546cc). A small-bodied representative ofthe Dmanisi paleodeme could havedispersed eastward across Wallace’sLine at 1.3-1.0 Ma. Once on Flores,its wrist and foot specialized to local

conditions. If a larger-bodied Dma-nisi hominin arrived at Flores in thesame time frame, island dwarfing isimplicated. In island contexts, ungu-lates can develop significantly short-ened limb bones, shortenedmetapodials, and stiffer joints, some-times resulting in fused elements.187

Moreover, small-bodied hominins(Australopithecus and Homo) tend tohave short, ape-like lower limbs as afunction of body size scaling.188

Currently, we conclude that LB1can best be seen as the product ofinsular effects on a representative ofthe Dmanisi paleodeme. Likely path-ways include the specialization ofsmall-bodied arrival or the dwarfingof a larger-bodied arrival. Eitherway, LB1’s small and specializedskeletal morphology reflects threefeatures of Pleistocene Flores fauna:phylogenetic continuity, low speciesrichness, and disharmony. “All threeaspects stem from the isolated posi-tion of the island and have resultedin the distinct morphological charac-teristics of the Flores fauna.”6

Modern Humans

The arrival date for Homo sapiensat ISEA and Sahul (New Guinea and

Australia) is coalescing on a thresh-old of 47 ka.189 ISEA’s two well-

dated early Homo sapiens sites fall

on or near the threshold: �47 ka at

the Tabon Caves (Palawan, Philip-

pines)97 and �42 ka at Niah Cave

(Borneo, Indonesia).190 Alternatively,Wadjak (Java, Indonesia), once con-

sidered the earliest ISEA modern

human, has been radiometrically

dated to 37–28 ka.191 Regarding

Sahul, a much larger area with many

more archeological sites, age deter-minations fall into a 47-40-ka inter-

val.189 The 47-ka arrival threshold

suggests dispersal out of Africa

toward the onset of MIS 3 (�59 ka),

when renewed warmth and wetnessmade the normally arid areas of

northwest Africa and Arabia more

habitable than during the preceding

MIS 4.192,193 Such conditions may

have initiated the primary H. sapiensout-of-Africa dispersal.194–196

Figure 9. Hypothetical ISEA hominin lineages based on fossil, paleogenomic, and archeological evidence: yellow 5 Homo sapiens;red 5 Homo neanderthalensis; blue 5 Homo erectus and derived lines. On the Homo erectus line, archeological evidence providesdivergence ages for Flores, Sulawesi, and Luzon early hominin lineages, presumably from Java origins. Green arrows show genomicintrogressions to dispersing Homo sapiens from Eurasian Neandertals and, possibly, from ISEA Homo erectus-derived lines. Since Homoheidelbergensis is considered a descendant of Homo erectus, we assume that the paleogenomic evidence carried by Homo erectus issimilar to that recorded recently at Sima de los Huesos in Spain.198 [Color figure can be viewed in the online issue, which is available atwileyonlinelibrary.com.]



At present, the latest ISEA earlyhominin fossil occurrences predatethe 47 ka H. sapiens arrival thresh-old: Ngandong Homo erectus at�143-130 ka,140,141,189 Liang BuaHomo floresiensis at �60 ka,50 andthe Callao hominin at 66.7 ka.8 Insum, current evidence does not sug-gest significant overlap betweenearly and modern hominins.

Denisovan Paleogenomics. Paleoge-nomics provides an emerging line ofevidence for understanding the rela-tionship between ISEA early and mod-ern hominins during the LatePleistocene (Fig. 9). Denisovan alleleshave been identified in widely situatedEurasian Pleistocene contexts: extremesouthwest and relatively early (Sima delos Huesos, 430 ka) and far northeastand much later (Denisova, 60ka).197–199 The Sima mitochondrialsequence is identified as Denisovanand is distinct from that of Neander-tals.197,199 The Denisovan mtDNAgenome shares, for at least one millionyears, a last common ancestor with theclade leading to Homo sapiens andNeandertals.197 The nuclear sequenceshows a closer relationship to Nean-dertals.200 The Denisovan genome isthus associated with several Eurasianearly hominin lineages.201 The genomealso shows evidence of another, evenolder hominin lineage.197,202

In modern human populations,Denisovan alleles appear in twowidely spaced areas of Eurasia:Tibet203 and east of Wallace’s Line,specifically in the Philippines andSahul islands of New Guinea andAustralia, and in Oceania.204,205

Given this far eastern distribution,Cooper and Stringer argue that Deni-sovan alleles introgressed into Homosapiens populations east of Wallace’sLine.204 Cooper and Stringer proposeHomo heidelbergensis as the east-dis-persing Denisovan-carrying earlyhominin. However, with no fossilevidence for this taxon in ISEA, theissue remains open. The ISEA LatePleistocene early hominin fossil can-didates for the introgression areNgandong Homo erectus, Homo flore-siensis, and the Luzon hominin, butit is not yet known if any of thesespecies carried Denisovan alleles.

CONCLUSION

Among all Old World paleoanthro-pological areas, ISEA is unique in itsgreat range of maritime environ-ments. Pleistocene glacial eustasygave the region a fast-changing char-acter during the early homininperiod. The current geography repre-sents an extreme sea-level highstandwith near maximum marine coverand terrestrial isolation. Such high-stands typified relatively short peri-ods of the Pleistocene. The generallylonger glacial lowstands, rangingfrom �20 m to �125 m below pres-ent sea level, produced more contin-uous terrestrial exposures and themeans for mammalian dispersal.Nevertheless, highstands were criti-cal for isolating large mammalianfauna, including early hominins.

The spatial context for ISEA earlyhominin biogeography centers onWallace’s Line. For more than twomillion years, glacial eustasy, tec-tonic uplift, and volcaniclastic depo-sition have structured andrestructured dispersal routes andprovince habitats east and west ofthe line. On Sunda, fossil evidencesuggests that early Homo erectusarrived from mainland Eurasia dur-ing a low sea-level stand before 1.6Ma (Solo Basin, Sangiran Forma-tion). On the Arc, stone tools indi-cate that hominins arrived at the Soa(Flores) and Walanae (South Sula-wesi) basins well before 1.0 Ma andto the Cagayan and Arubo basins(Luzon) probably by 800 ka.

The evidence of interglacial isola-tion is more subtle. By 300 ka, Sundagroups inhabited caves and used flaketools to process rhinoceros, tapir, andcervids (Gunung Sewu karst). LiangBua cave opened for hominin habita-tion by �195 ka, when the Floreshominin used an advanced flake toolindustry in pursuit of the islands’endemic large fauna (Wae Racangkarst). Most of the later evidencepostdates the Last Interglacial (MIS5e). By this time, morphological spe-cializations include the large-brainedNgandong Homo erectus and thesmall Homo floresiensis body size andbrain, both of which suggest thatselection for these features developedbefore MIS 5e.

Liang Bua skeletal morphology likely

represents development in relation tolow rates of resource availability and

predator stress. The diversification anddwarfing of ISEA Stegodontidae pro-

vides a relevant large mammalian fossilanalogue. Liang Bua archeology indi-cates social and technological organi-

zation within which a very smallhuman could consume large verte-

brates. Parallels may be drawn withLate Pleistocene western Eurasia,

where Homo neanderthalensis showedspecializations in skeletal morphology

and stone technology. The Last Inter-glacial may have isolated ISEA early

hominin and stegodon populations totheir evolutionary limits. The early

hominin period apparently ended withthe arrival of Homo sapiens.

ACKNOWLEDGMENTS

This long-term project was funded

by the L.S.B. Leakey Foundation (4grants), the Wenner- Gren Foundation

for Anthropological Research (2grants), and the National Science

Foundation (2 grants). Continuing sup-port has also come from the Human

Evolution Research Fund at the Uni-versity of Iowa Foundation. New fund-

ing from the Center for Global andRegional Environmental Research

(CGRER) at The University of Iowaand the Ann and Gordon Getty Foun-

dation is being used to continue andexpand this research. For research

assistance, we thank K. Lindsay Eaves,Natalie O’Shea, Kiran Patel, Toby Ava-los, Chloe Daniel, and Madeleine Hoof-

nagle. Gregg Gunnell and JonathanBloch carefully reviewed the manu-

script. For our many years of field col-laboration, we gratefully acknowledge

Yahdi Zalm, Yan Rizal, and A. Aswan(Institute of Technology-Bandung,

Java), E. Arthur Bettis Ill (University ofIowa), Gregg Gunnell (Duke Univer-

sity), Frank Huffman (University ofTexas, Austin), and John-Paul Zonne-

veld (University of Alberta).

REFERENCES

1 Wallace AR. 1860. On the zoological geogra-phy of the Malay Archipelago. Zool J Linn SocLond 4:172–184.

2 Dubois E. 1892. Palaeontologische onderzoe-kingen op Java: Verslang van het Mijnwezen 3:10–14.


3 Lordkipanidze D, Jashashvili T, Vekua A,et al. 2007. Postcranial evidence from earlyHomo from Dmanisi, Georgia. Nature 449:305–310.

4 Brown P, Sutikna T, Morwood M, et al. 2004.A new small-bodied hominin from the LatePleistocene of Flores, Indonesia. Nature 431:1055–1061.

5 Morwood MJ, Soejono RP, Roberts RA, et al.2004. Archaeology and age of a new homininfrom Flores in eastern Indonesia. Nature 431:1087–1091.

6 Meijer HJM, van den Hoek Ostende LW, vanden Bergh GD, et al. 2010. The fellowship ofthe hobbit: the fauna surrounding Homo flore-siensis. J Biogeogr 37:995–1006.

7 Dennell RW, Louys J, O’Regan HJ, et al.2014. The origins and persistence of Homo flor-esiensis on Flores: biogeographical and ecologi-cal perspectives. Quaternary Sci Rev 96:98–107.

8 Mijares AS, Detroit F, Piper P, et al. 2010.New evidence for a 67,000-year-old humanpresence at Callao Cave, Luzon, Philippines.J Hum Evol 59:123–132.

9 Mijares A, personal communication (e-mail,November 11, 2014).

10 Hays JD, Imbrie J, Shackleton NJ. 1976.Variations in the earth’s orbit: pacemaker ofthe Ice Ages. Science 194:1121–1132.

11 Raymo ME, Nisancioglu K. 2003. The 41kyr world: Milankovitch’s other unsolved mys-tery. Paleoceanography 18:1011–1016.

12 Mudelsee M, Schulz M. 1997. The Mid-Pleistocene climate transition: onset of 100 kacycle lags ice volume build-up by 280 ka. EarthPlanet Sci Lett 151:117–123.

13 Lourens LJ, Antonarakou A, Hilgen FJ, et al.1996. Evaluation of the Plio-Pleistocene astro-nomical timescale. Paleoceanography 11:391–413.

14 Joordens JCA, Dupont-Nivet G, Feibel CS,et al. 2013. Improved age control on earlyHomo fossils from the upper Burgi Member atKoobi Fora, Kenya. J Hum Evol 65:731–745.

15 Zhisheng A, Clemens SC, Shen J, et al.2011. Glacial-interglacial indian summer mon-soon dynamics. Science 333:719–723.

16 Shackelton NJ. 1997. The deep-sea sedimentrecord and the Pliocene-Pleistocene boundary.Quaternary Int 40:33–35.

17 Hall R. 2001. Cenozoic reconstructions ofSE Asia and the SW Pacific: changing patternsof land and sea. In: Metcalfe I, Smith JMB,Morwood M, et al., editors. Faunal and floralmigrations and evolution in SE Asia-Austral-asia. Lisse: Swets and Zeitlinger. p 35–56.

18 de Deckker P, Tapper NJ, van der Kaars S.2002. The status of the Indo-Pacific Warm Pooland adjacent land at the Last Glacial Maxi-mum. Global Planet Change 35:25–35.

19 Larick R, Ciochon RL, Zaim Y, et al. 2001.Early Pleistocene 40Ar/39Ar ages for BapangFormation hominins, Central Jawa, Indonesia.Proc Natl Acad Sci USA 98:4866–4871.

20 Larick R, Ciochon RL, Zaim Y, et al. 2000.Lithostratigraphic context for Kln-1993.05-SNJ,a fossil colobine maxilla from Jokotingkir, San-giran Dome. Int J Primatol 21:731–759.

21 Watanade N, Kadar D, editors. 1985. Qua-ternary geology of the hominid fossil bearingformations in Java. Bandung: GeologicalResearch and Development Centre.

22 Zaim Y, Larick R, Ciochon RL, et al. 1999.Karakteristik satuan Lahar Bawah dari formasiPucangan di Sangiran, Jawa Tengah. BuletinGeolgi Institut Teknologi Bandung 31:67–84.

23 Bettis E, Zaim Y, Larick R, et al. 2004.Landscape development preceding Homo erec-tus immigration into Central Java, Indonesia:the Sangiran Formation Lower Lahar. Palaeo-geogr Palaeoclimatol Palaeoecol 206:115–131.

24 S�emah AM. 1984. Palynology and JavanesePithecanthropus environment. Courier For-shungsinstitut Senckenberg 69:237–243.

25 Tonkunaga S, Oshima H, Polhaupesy A,et al. 1985. A palynological study of the Pucan-gan and Kabuh formations in the Sangiranarea. In: Watanabe N, Kadar D, editors. Qua-ternary geology of the hominid fossil bearingformations in Java. Bandung: GeologicalResearch and Development Centre. p 199–217.

26 de Vos J, Sondaar PY, van den Bergh GD,et al. 1994. The Homo bearing deposits of Javaand its ecological context. Courier Forshungsin-stitut Senckenberg 171:129–140.

27 Aimi M, Aziz F. 1985. Vertebrate fossilsfrom the Sangiran dome, Mojokerto, Trinil andSambungmacan Areas. In: Watanabe N, KadarD, editors. Quaternary geology of the hominidfossil bearing formations in Java. Bandung:Geological Research and Development Centre.p 155–197.

28 Bouteaux A. 2005. Pal�eontologie,pal�eo�ecologie et taphonomie des mammiferesdu Pl�eistocene moyen ancien du site �a homi-nid�es de Sangiran (Java central, Indon�esie).Ph.D. thesis, Mus�eum National d’HistoireNaturelle Paris.

29 Bettis EA, Milius A, Carpenter S, et al. 2009.Way out of Africa: Early Pleistocene paleoenvir-onments inhabited by Homo erectus in San-giran, Java. J Hum Evol 56:11–24.

30 Rozzi R, Winkler DE, De Vos J, et al. 2013.The enigmatic bovid Duboisia santeng (Dubois,1891) from the Early-Middle Pleistocene ofJava: a multiproxy approach to its paleoecol-ogy. Palaeogeogr Palaeoclimatol Palaeoecol377:73–85.

31 Choi K, Driwantoro D. 2007. Shell tool useby early members of Homo erectus in Sangiran,central Java, Indonesia: cut mark evidence.J Archaeol Sci 34:48–58.

32 van den Bergh, G. D. 1999. The Late Neo-gene elephantoid-bearing faunas of Indonesiaand their palaeozoogeographic implications: astudy of the terrestrial faunal succession ofSulawesi, Flores and Java, including evidencefor early hominid dispersal east of Wallace’sLine. Scripta Geol 117:1–419.

33 van den Bergh GD, de Vos J, Aziz F, et al.1999. Homo erectus in S.E. Asia: time, spaceand migration routes, I. The Flores case. In:Gilbert J, Sanchez F, Gilbert L, et al., editors.The hominids and their environment duringthe Lower and Middle Pleistocene of Eurasia:Proceedings of the International Conference ofHuman Paleontology. Orce: Museo de Prehisto-ria y Paleontologia. p 353–362.

34 Dubois E. 1894. Pithecanthropus erectus einemenschen€ahnliche €Ubergangsform aus Java.Batavia: Landesdruckerei.

35 Dubois E. 1908. Das geologische alter derKendeng-oder Trinil-fauna. Tijdschr Kon NedAardr Gen 25:1235–1270.

36 Dubois E. 1932. The distinct organization ofPithecanthropus of which the femur bears evi-dence, now confirmed from other individualsof the described species. Proc KoninklijkeNederlandse Akademie van Wetenschappen 35:716–722.

37 Dubois E. 1934. New evidence of the dis-tinct organization of Pithecanthropus. Proc.Koninklijke Nederlandse Akademie van Weten-schappen 37: 139–145.

38 Dubois E. 1935. The sixth (fifth new) femurof Pithecanthropus erectus. Proc KoninklijkeNederlandse Akademie van Wetenschappen 38:850–852.

39 Barstra G-J. 1982. The river-laid strata nearTrinil, site of Homo erectus, Java, Indonesia. In:Bartstra GJ, Casparie WA, editors. Modern qua-ternary research in Southeast Asia, vol. 7. Rot-terdam: Balkema. p 97–130.

40 Bartsiokas A, Day MH. 1993. Electron-probeenergy-dispersive x-ray-microanalysis (edxa) inthe investigation of fossil bone: the case of JavaMan. Proc R Soc B Lond 252:115–123.

41 Ruff CB, Puymerail L, Macchiarelli R, et al.2015. Structure and composition of the Trinilfemora: functional and taxonomic implications.J Hum Evol 80:147–158.

42 Selenka ML, Blanckenhorn M. 1911. DiePithecanthropus-Schichten auf Java, geologi-sche und pal€aontologische ergebnisse derTrinil-Expedition (1907 und 1908). Leipzig:Wilhelm Engelmann.

43 de Vos J, Aziz F. 1989. The excavations byDubois (1891–1900), Selenka (1906–1908), andthe Geological Survey by the Indonesian-Japanese Team (1976–1977) at Trinil Java,Indonesia. J Anthropol Soc Nip 97:407–420.

44 de Vos J, Sartono S, Hardja-Sasmita S,Sondaar PY. 1982. The fauna from Trinil, typelocality of Homo erectus; a reinterpretation.Geol en Mijnbouw 61:207–211.

45 Storm P. 2012. A carnivorous niche for JavaMan? A preliminary consideration of the abun-dance of fossils in Middle Pleistocene Java.Comptes Rendus Palevol 11:191–202.

46 Joordens JCA, d’Errico F, Wesslingh FP,et al. 2014. Homo erectus at Trinil on Java usedshells for tool production and engraving.Nature 518: 228–231.

47 Brumm A, Jensen G, van den Bergh G, et al.2010. Hominins on Flores, Indonesia, by onemillion years ago. Nature 464:748–752.

48 Aziz F, van den Bergh GD, Morwood MJ,et al. 2009. Excavations at Tangi Talo, centralFlores, Indonesia. In: Aziz F, Morwood MJ, vanden Bergh GD, editors. Palaeontology andarchaeology of the Soa Basin, Central Flores,Indonesia. Bandung: Indonesian GeologicalSurvey Institute. p 41–58.

49 van den Bergh G, personal communication(e-mail, September 15, 2014).

50 Brumm A, personal communication (e-mail,February 24, 2013).

51 Hooijer DA. 1954. Pleistocene vertebratesfrom Celebes VIII: dentiton and skeleton of Cel-ebochoerus heekereni Hooijer. Zool Verhande-lingen 24:1–44.

52 van Heekeren HR. 1972. The Stone Age ofIndonesia. Verhandeligngen 61. The Hague:Martinus Nithoff.

53 Hooijer DA, Bartstra GJ. 1995. Fossils andartefacts from South-western Sulawesi (Cele-bes). In: Bower JRF, Sartono S, editors. Evolu-tion and ecology of Homo erectus. Leiden:Pithecanthropus Centennial Foundation. p 79–82.

54 Keates SG, Bartstra GJ. 2000. Observationon Cabengian and Pacitanian artifacts fromisland Southeast Asia. Quart€ar 51/52: 9–32.

55 Corvinus G. 2004. Homo erectus in East andSoutheast Asia, and the questions of the age ofthe species and its association with stone arti-facts, with special attention to handaxe-liketools. Quaternary International 117: 141–151.

56 Bartstra GJ. 1977. Walanae formation andWalanae terraces in the stratigraphy of South


Sulawesi (Celebes, Indonesia). Quartaer 27/28:21–30.

57 McKie R. 2010. How a hobbit is rewritingthe history of the human race. The Observer,20 February 2010. http://www.theguardian.com/science/2010/feb/21/hobbit-rewriting-history-human-race

58 Berger WH, Bickert T, Schmidt H, Wefer G.1993. Quaternary oxygen isotope record ofpelagic foraminifers: site 806, Ontong Java Pla-teau. Proc ODP, Sci Results 130:381–395.

59 Imbrie J, Berger A, Boyle EA, et al. 1993.On the structure and origin of major glaciationcycles 2: the 100,000-year cycle. Paleoceanogra-phy 8: 699–735.

60 Muttoni G, Scardia G, Kent DV. 2010.Human migration into Europe during the lateEarly Pleistocene climate transition. Palaeo-geography, Palaeoclimatology, Palaeoecology296: 79–93.

61 Muttoni G, Scardia G, Kent DV, et al. 2011.First dated human occupation of Italy at�0.85 Ma during the late Early Pleistocene cli-mate transition. Earth Planet Sci Lett 307:241–252.

62 Clark PU, Archer D, Pollard D, et al. 2006.The middle Pleistocene transition: characteris-tics, mechanisms, and implications for long-term changes in atmospheric pCO2. Q Sci Rev25:3150–3184.

63 Pena L, Goldstein S. 2014. Thermohalinecirculation crisis and impacts during the mid-Pleistocene transition. Science 345: 318–322.

64 Head MJ, Gibbard PL. 2005. Early-MiddlePleistocene transitions: an overview and recom-mendation for the defining boundary. In: HeadMJ, Gibbard PL, editors. Early-Middle Pleisto-cene transitions: the land–ocean evidence. Lon-don: Geological Society, Special Publications247. p 1–18.

65 Berger WH, Jansen E. 1994. Mid-Pleistoceneclimate shift: the Nansen connection. In:Johannessen OM, Muench RD, Overland JE,editors. The Polar oceans and their role inshaping the global environment. Washington,DC: American Geophysical Union. p 295–311.

66 S�emah F, S�emah AM, Djubiantono T, et al.1992. Did they also make stone tools? J HumEvol 23:439–446.

67 Saleki H. 1997. Apport d’une intercomparai-son des m�ethodes nucl�eaires (230Th/234U, ESRet 40Ar/39Ar) �a la datation de couches fossili-feres pleistocenes dans le dome de Sangiran(Java, Indon�esie). Ph.D. thesis, Mus�eumNational d’Histoire Naturelle Paris.

68 Unpublished 40Ar/39Ar date from Iowa-Bandung Java Project, Fall 2006.

69 Bouteaux A, Moigne A-M. 2010. New tapho-nomical approaches: The Javanese Pleistoceneopen-air sites (Sangiran, Central Java). Q Int223-224:220–225.

70 S�emah AM, S�emah F, Moudrikah R, et al.2004. A Late Pleistocene and Holocene sedimen-tary record in Central Java and its palaeocli-matic significance. In: Keates SG, Pasveer JM,editors. Modern quaternary research in South-east Asia, vol. 18. London: CRC Press. p 63–88.

71 S�emah F. 2001. La position stratigraphiquedu site de Ngebung 2 (dome de Sangiran, JavaCentral, Indon�esie). In: S�emah F, Falgueres C,Grimaud-Herv�e D, et al., editors. Origine desPeuplements et Chronologie des CulturePal�eolithiques dans le Sud-Est Asiatique. Paris:Semenanjung. p 299–377.

72 S�emah F, S�emah AM, Simanjuntak T. 2003.More than a million years of human occupa-tion in insular Southeast Asia: the early archae-ology of eastern and Central Java. In: Mercader

J, editor. Under the canopy: the archaeology oftropical rainforests. New Brunswick: RutgersUniversity Press. p 161–190.

73 Sharon G. 2007. Acheulian large flakeindustries: technology, chronology, and signifi-cance. Oxford: Archaeopress.

74 Sharon G. 2010. Large flake Acheulian. QInt 223-224:226–233.

75 Bartstra GJ. 1984. Dating the Pacitanian:some thoughts. Courier Forschungs Inst Sen-kenberg 69:253–258.

76 Sondaar PY, Aziz F, de Vos J, et al. 1994.Middle Pleistocene faunal turnover and coloni-zation of Flores (Indonesia) by Homo erectus.Comptes Rendus Acad Sci 319:1255–1262.

77 van den Bergh GD, de Vos J, Aziz F, et al.2001. Elephantoidea in the Indonesian region:new Stegodon findings from Flores. In: Theworld of elephants. Rome: Proceedings of theFirst International Congress. p 623–627.

78 Brumm A, Moore M, van den Bergh G,et al. 2010. Stone technology at the MiddlePleistocene site of Mata Menge, Flores, Indone-sia. J Archaeol Sci 37:451–473.

79 Verhoeven T. 1959. Die Klingenkultur derInsel Timor. Anthropos 54:970–972.

80 Hooijer DA. 1972. Stegodon trigoncephalusflorensis Hooijer and Stegon timorensis Sartonofrom the Pleistocene of Flores and Timor. ProcKoninklijke Nederlandsche Akad Wetenschap-pen Series B 75:27–33.

81 Hooijer DA. 1972. Varanus (Repitlia, Sauria)from the Pleistocene of Timor. Zool Mededlin-gen 47:445–448.

82 Verhoeven T. 1964. Stegodon-Fossilien aufder Insel Timor. Anthropos 59: 634.

83 Hooijer DA. 1969. The Stegodon fromTimor. Proc Koninklijke Nederlandse AkadWetenschappen Amsterdam B 72:203–210.

84 Sartono S. 1969. On the Plio-Pleistoceneboundary of Java. Bulletin National Institute ofGeology and Mining (Bandung) 2: 1–19.

85 Sartono S. 1973. On Pleistocene migrationroutes of vertebrate fauna in Southeast Asia.Geol Soc Malaysia 6:273–286.

86 Sartono S. 1969. Stegodon timorensis: apygmy species from Timor (Indonesia). ProcKoninklijke Nederlandse Akad WetenschappenAmsterdam B 72:192–202.

87 Bednarik RG. 2000. Pleistocene Timor:some corrections. Aust Archaeol 51:16–20.

88 Soejono RP. 1982. Data Baru TentangIndustri Paleolitik di Indonesia. In: PertemuanIlmiah Arkeologi – KE II. Jakarta: Pusat Peneli-tian Arkeologi Nasional. p 33–45.

89 Sartono S. 1979. The discovery of a pygmystegodon from Sumba, East Indonesia: anannouncement. In: Bartstra GJ, Casparie WA,editors. Modern quaternary research in South-east Asia, vol. 5. Rotterdam: Balkema. p 57–63.

90 Hooijer DA. 1981. What, if anything new, isStegodon sumbaensis Sartono? In: Bartstra GJ,Casparie WA, editors. Modern quaternaryresearch in Southeast Asia, vol. 6. Rotterdam:Balkema. p 89–90.

91 Setiyabudi E, Kurniawan I, van den BerghG. 2012. Fossils of Stegodon and Varanus komo-doensis Sumba and Flores: a Pleistocene land-bridge. IAGI Digital Proceedings of 2012Annual Meeting. Indonesia: IAGI. p 1–1.

92 Beyer H. 1956. New finds of fossil mammalsfrom Pleistocene strata of the Philippines. Que-zon City: National Research Council of thePhilippines.

93 von Koenigswald, GHR. 1956. Fossil mam-mals from the Philippines. Quezon City:National Research Council of the Philippines.

94 Fox RB. 1978. The Philippine Paleolithic. In:Ikawa-Smith F, editor. Early Paleolithic inSouth and East Asia. The Hague: Mouton Pub-lishers. p 59–85.

95 Bautista AP. 1991. Recent zooarchaeologicalresearches in the Philippines. J Arkeol 4:45–58.

96 de Vos J, Bautista A. 2001. An update onthe vertebrate fossils from the Philippines. NatlMuseum Papers 11:62–105.

97 Dizon E, Pawlik A. 2010. The lower Palaeo-lithic record in the Philippines. Q Int 223-224:444–450.

98 Mathisen ME, Vondra CF. 1983. The fluvialand pyroclastic deposits of the Cagayan Basin,northern Luzon, Philippines: an example ofnon-marine volcaniclastic sedimentation in aninterarc basin. Sedimentology 30:369–392.

99 Wasson RJ, Cochrane RM. 1979. Geologicaland geomorphological perspectives on archaeo-logical sites in the Cagayan valley, NorthernLuzon, the Philippines. In: Bartstra G-J, Cas-parie WA, editors. Modern quaternary researchin Southeast Asia, vol. 5. Rotterdam: Balkema.p 1–26.

100 de Vos J. Personal communication (e-mail,August 4, 2014)

101 Jensen GM, Storey M, Roberts RG, et al.2010. Combined 40Ar/39Ar and OSL dating ofPleistocene pyroclastic and fluvial deposits inthe Cagayan Valley Basin, Northern Luzon,Philippines. Geophys Res Abst 12: EGU2010-3937-2.

102 Dizon E. 1998. Origins of people. In: Mag-daraog G, editor. Environment and naturalresources atlas of the Philippines. Quezon City:Environmental Center of the Philippines Foun-dation. p 70–91.

103 Pawlik AF. 2002. Acheulean in NuevaEcija? Hukay: J University Philippines ArchaeolStud Program 4:1–22.

104 Teodosio S. 2006. A functional analysis ofthe Arubo stone tools. Master’s thesis, Univer-sity of the Philippines.

105 Jansen JHF, Kuijpers A, Troelstra SR.1986. A Mid-Brunhes climatic event: long-termchanges in global atmosphere and ocean circu-lation. Science 232:619– 622.

106 Augustin L, Barbante C, Barnes PR, et al.2004. Eight glacial cycles from an Antarctic icecore. Nature 429:623–628.

107 Chappell J, Shackleton NJ. 1986. Oxygenisotopes and sea level. Nature 324:137–140.

108 Tzedakis PC, Hooghiemstra H, P€alike H.2006. The last 1.35 million years at TenaghiPhilippon: revised chronostratigraphy and long-term vegetation trends. Q Sci Rev 25:3416–3430.

109 Berger WH, Wefer G. 2003. On the dynam-ics of the ice ages: Stage-11 paradox, mid-Brunhes climate shift and 100-ky cycle. In:Droxler AW, Poore RZ, Burckle LH, editors.Earth’s climate and orbital eccentricity: theMarine Isotope Stage 11 question. Washington,DC: American Geophysical Union. p 41–59.

110 Hearty PJ, Kinder P, Cheng H, et al. 1999.A 120 m middle Pleistocene sea-level highstand(Bermuda and the Bahamas) due to partial col-lapse of Antarctic ice. Geology 27:375–378.

111 Stringer C. 2012. The status of Homo hei-delbergensis (Schoetensack 1908). Evol Anthro-pol 21:101–107.

112 Wang YJ, Cheng H, Edwards RL, et al.2008. Millennial- and orbital-scale changes inthe East Asian monsoon over the past 224,000years. Nature 451:1090–1093.

113 Cheng H, Edwards RL, Wang Y, et al.2006. A penultimate glacial monsoon record


http://www.theguardian.com/science/2010/feb/21/hobbit-rewriting-history-human-race



from Hulu Cave and two-phase glacial termina-tions. Geology 34:217–220.

114 Wang YJ, Cheng H, Edwards RL, et al.2001. A high-resolution absolute-dated latePleistocene monsoon record from Hulu Cave,China. Science 294:2345–2348.

115 Cheng H, Edwards RL, Broecker WS, et al.2009. Ice Age terminations. Science 326:248–252.

116 Edwards RL, personal communication.National Academy of Sciences Day Prize Lec-ture at Oberlin College, Nov. 21, 2013.

117 Westaway KE, Morwood MJ, Roberts RG,et al. 2007. Age and biostratigraphic signifi-cance of the Punung Rainforest Fauna, EastJava, Indonesia, and implications for Pongoand Homo. J Hum Evol 53:709–717.

118 Lehmann H. 1936. Morphological studiesin Java. In: Sweeting MM, editor. Karst geo-morphology. Stroudsburg, PA: HutchinsonRoss. p 320–328.

119 van Bemmelen RW. 1949. The geology ofIndonesia. In: van Bemmelen RW, editor. Gen-eral geology of Indonesia and adjacent archipe-lagoes, vol. 1A. The Hague: GovernmentPrinting Office.

120 Semah F, Semah AM, Djubiantono T.1998. From the shoreline to the slopes of thevolcanoes: the long Pithecanthropus trek. In:Simanjuntak T, Prasetyo B, Handini R, editors.Sangiran: man, culture and environment inPleistocene times. Jakarta: National ResearchCentre of Archaeology and Ecole Francaised’Extreme Orient, Yayasan Obor Indonesia. p195–218.

121 Balter M. Etchings on a 500,000-year-oldshell appear to have been made by humanancestor. Science. Online news article dated 3December 2014. http://news.sciencemag.org/archaeology/2014/12/etchings-500000-year-old-shell-appear-have-been-made-human-ancestor

122 Henshilwood CS, d’Erico F, van NiekerkKL, et al. 2011. A 100,000-year-old ochre-processing workshop at Blombos Cave, SouthAfrica. Science 334:219–222.

123 Marean CW, Bar-Matthews M, BernatchezJ, et al. 2007. Early human use of marineresources and pigment in South Africa duringthe Middle Pleistocene. Nature 449:905–909.

124 Simanjuntak T. 2004. New insight on theprehistoric chronology of Gunung Sewu, Java,Indonesia. In: Keates SG, Pasveer JM, editors.Modern quaternary research in Southeast Asia,vol. 18. London: CRC Press. p 9–30.

125 Bartstra G-J. 1976. Contributions to thestudy of the Palaeolithic Patjitan culture, Java,Indonesia. Leiden: E. J. Brill.

126 S�emah F, S�emah AM. 2006. Palaeolithicsettlements in the Southeast Asian archipela-gos: an Indonesian perspective. In: Simanjun-tak T, Hisyam M, Prasetyo B, et al., editors.Archaeology: Indonesian perspective: R.P. Soe-jono Festschrift. Jakarta: Indonesian Instituteof Sciences. p 148–161.

127 Hameau S. 2004. Datation despal�eolithques du Pl�eistocene moyen et superiorde Punung (Pacitan, Java) application de meth-ods ESR et U/Th aux grottes du Song Terus etGoa Tabuhan. Ph.D. thesis, Mus�eum Nationald’Histoire Naturalle Paris.

128 Weidenreich F. 1951. Morphology of SoloMan. Anthropol Papers Am Museum Nat Hist43:205–290.

129 Bartstra GJ, Soegondho S, van der Wijk A.1988. Ngandong man: age and artifacts. J HumEvol 17:325–337.

130 Huffman OF, de Vos J, Berkhout AW,et al. 2010. Provenience reassessment of the

1931-1933 Ngandong Homo erectus (Java), con-firmation of the bone-bed origin reported bythe discoverers. PaleoAnthropology 2010:1–60.

131 Schwartz JH, Tatersall I. 2005. The humanfossil record, vol 4: craniodental morphology ofearly hominids (genera Australopithecus, Para-nthropus, Orrorin), and overview. Hoboken, NJ:Wiley-Liss.

132 Delson E, Harvati K, Reddy D, et al. 2001.The Sambungmacan 3 Homo erectus calvaria:a comparative morphometric and morphologi-cal analysis. The Anatomical Record 262:380–397.

133 M�arquez S, Mowbray K, Sawyer GJ, et al.2001. New fossil hominid calvaria fromIndonesia-Sambungmacan 3. Anat Rec 262:344–368.

134 Huffman OF, Balzeau A, Berkhout AW,et al. 2010. Mass death and lahars in thetaphonomy of the Ngandong Homo erectusbonebed, and volcanism in the hominin recordof Eastern Java. PaleoAnthropology 2010:A14.

135 Sondaar PY. 1984. Faunal evolution andthe mammalian biostratigraphy of Java. CourForsch-Inst Senckenberg 69:219–235.

136 Ciochon RL, Huffman OF, Bettis III EA,et al. 2009. Rediscovery of the Homo erectusbed at Ngandong: site formation of a late Pleis-tocene hominin site in Asia. Am J Phys Anthro-pol 48(suppl):110.

137 von Koenigswald GHR. 1939. DasPleistoc€an Javas. Quart€ar Jahrbuch f€ur dieErforschung des Eiszeitalters und seiner Kultu-ren Berlin 2:28–53.

138 Swisher CC, Rink WJ, Ant�on SC, et al.1996. Latest Homo erectus of Java: Potentialcontemporaneity with Homo sapiens in South-east Asia. Science 274:1870–1874.

139 Yokoyama Y, Falgueres C, S�emah F, et al.2008. Gamma-ray spectrometric dating of lateHomo erectus skulls from Ngandong and Sam-bungmacan, Central Java, Indonesia. Journal ofHuman Evolution 55:274–277.

140 Indriati E, Swisher III CC, Lepre C, et al.2011. The age of the 20 meter Solo River Ter-race, Java, Indonesia and the survival of Homoerectus in Asia. PLoS ONE 6:e21562. doi:10.1371/journal.pone.0021562

141 Van Den Bergh GD, Westaway K,Kurniawan I, et al. 2014. A Late Pleistoceneflaking surface from the 20-m Solo River Ter-race, Central Java, Indonesia. In: Abstracts ofthe 20th Indo-Pacific Prehistory AssociationCongress, Siem Reap, Cambodia. p 261.

142 Rasmussen M, Guo X, Wang Y, et al. 2011.An Aboriginal Australian genome reveals separatehuman dispersals into Asia. Science 334:94–98.

143 Collings HD. 1938. Pleistocene site in theMalay Peninsula. Nature 142:575–576.

144 Movius HL Jr. 1949. Lower Paleolithicarchaeology in southern Asia and the Far East.In: Howells WW, editor. Early man in the FarEast: studies in physical anthropology 1.Detroit: American Association of PhysicalAnthropologists. p 17–77.

145 Harrison T. 1975. Tampan: Malaysia’sPalaeolithic reconsidered. In: Bartstra G-J, Cas-parie WA, editors. Modern quaternary researchin Southeast Asia, vol. 1. Rotterdam: Balkema.p 53–70.

146 Zuraina M, Tjia HD. 1988. Kota Tampan,Perak: the geological and archaeological evi-dence for a late Pleistocene site. Journal of theMalaysian Branch of the Royal Asiatic Society61: 123–134.

147 Westgate JA, Shane P, Pearce N, et al.1998. All Toba tephra occurences across penin-

sular India belong to the 75,000 year bp erup-tion. Q Res 50:107–112.

148 Ninkovich D, Sparks RSJ, Ledbetter MT.1978. The exceptional magnitude and intensityof the Toba eruption, Sumatra: an example ofthe use of deep-sea tephra layers as a geologicaltool. B Volcanol 41:286–298.

149 Chesner CA, Rose WI, Deino A, et al. 1991.Eruptive history of earth’s largest Quaternarycaldera (Toba, Indonesia) clarified. Geology 19:200–203.

150 Zuraina M. 1998. Radiocarbon dates andculture sequence in the Lenggong Valley andbeyond. Malaysia Museums J 34:241–249.

151 Oppenheimer S. 2012. A single southernexit of modern humans from Africa: before orafter Toba? Q Int 258:88–99.

152 Westaway KE, Sutikna T, Saptomo WE,et al. 2009. Reconstructing the geomorphic his-tory of Liang Bua, Flores, Indonesia: a strati-graphic interpretation of the occupationalenvironment. J Hum Evol 57:465–483.

153 van den Bergh GD, Due Awe RA, MorwoodMJ, et al. 2008. The youngest Stegodon remainsin Southeast Asia from the Late Pleistocenearchaeological site Liang Bua, Flores, Indone-sia. Quatern Int 182: 16–48.

154 Brumm A, Aziz F, van den Bergh GD,et al. 2006. Early stone technology on Floresand its implications for Homo floresiensis.Nature 441:624–628.

155 van den Bergh GD, Meijer HJM, Awe R,et al. 2009. The Liang Bua faunal remains: a 95kyr sequence from Flores, East Indonesia.J Hum Evol 57:527–537.

156 Mijares ASB. 2005. The archaeology ofPenablanca cave sites, northern Luzon, Philip-pines. J Austronesian Stud 1:65–93.

157 McDonald RC. 1976. Limestone morphol-ogy in South Sulawesi, Indonesia. ZeitschrGeomorphol 26(suppl):79–91.

158 Bulbeck D, Sumantri I, Hiscock P, et al.2004. Leang Sakapao 1, a second dated Pleisto-cene site from South Sulawesi, Indonesia. In:Keates SG, Pasveer JM, editors. Modern quater-nary research in Southeast Asia, vol. 18. Lon-don: CRC Press. p 111–128.

159 Glover IC. 1981. Leang Burung 2: anUpper Palaeolithic rock shelter in south Sula-wesi, Indonesia. In: Bartstra GJ, Casparie WA,editors. Modern quaternary research in South-east Asia, vol. 6. Rotterdam: Balkema. p 1–38.

160 Aubert M, Brumm A, Ramli, et al. 2014.Pleistocene cave art from Sulawesi, Indonesia.Nature 514:223–227.

161 Brumm A, personal communication (e-mail, February 1, 2015).

162 Harrison T, Krigbaum J, Manser J. 2006.Primate biogeography on the Sunda ShelfIslands: a paleontological and zooarchaeologi-cal perspective. In: Lehman SM, Fleagle JG,editors. Primate biogeography. New York:Springer. p 331–372.

163 Birdsell JH. 1977. The recalibration of aparadigm for the first peopling of greater Aus-tralia. In: Allen J, Golson J, Jones J, editors.Sunda and Sahul: prehistoric studies in South-east Asia, Melanesia and Australia. London:Academic Press. p 113–168.

164 Voris HK. 2000. Maps of Pleistocene sealevels in Southeast Asia: shorelines, river sys-tems and time durations. J Biogeogr 27:1153–1167.

165 Wallace AR. 1876. The geographical distri-bution of animals. London: Macmillan.

166 Sprintall J, Gordon AL, Koch-Larrouy A.et al. 2014. The Indonesian seas and their role


http://news.sciencemag.org/archaeology/2014/12/etchings-500000-year-old-shell-appear-have-been-made-human-ancestor



info:doi/10.1371/journal.pone.0021562

in the coupled ocean–climate system. NatureGeosci 7:487–492.

167 Morwood MJ, Aziz F, O’Sullivan P, et al.1998. Fission-track age of stone tools and fos-sils on the east Indonesian island of Flores.Nature 392:173–176.

168 Wallace AR. 1881. Island life or the phe-nomena and causes of insular faunas and flo-ras. New York: Harper and Brothers.

169 Howe HF, Smallwood J. 1982. Ecology ofseed dispersal. Annu Rev Ecol Syst 13:201–228.

170 Lazell JD. 2005. Island: fact and theory innature. Oakland: University of California Press.

171 Linkem CW, Brown RM, Siler CD, et al.2013. Stochastic faunal exchanges drive diversi-fication in widespread Wallacean and Pacificisland lizards (Squamata: Scincidae: Lamprole-pis smaragdina). J Biogeogr 40:507–520.

172 Saegusa H. 1996. Stegodontidae: evolutionaryrelationships. In: Shoshani J, Tassy P, editors. TheProboscidea: evolution and palaeoecology of ele-phants and their relatives. Oxford: Oxford Univer-sity Press. p 178–190.

173 Saegusa H, Thasod Y, Ratanasthien B.2005. Notes on Asian stegodontids. Q Int 126-128:31–48.

174 Hooijer DA. 1957. A Stegodon from Flores.Treubia 24:119–129.

175 van den Berg GD, Sondaar PY, de Vos J,Aziz F. 1996. The proboscideans of the South-East Asian islands. In: Shoshani J, Tassy P, edi-tors. The Proboscidea: evolution and palaeoe-cology of elephants and their relatives. Oxford:Oxford University Press. p 240–248.

176 Dennell RW, Louys J, O’Regan HJ, et al.2013. The origins and persistence of Homo flor-esiensis on Flores: biogeographical and ecologi-cal perspectives. Q Sci Rev 96:98–107.

177 Muttoni G, Kent DV, Scardia G, et al.2015. Bottleneck at Jaramillo for human migra-tion to Iberia and the rest of Europe? J HumEvol 80:187–190.

178 Martin RD, MacLarnon AM, Phillips JL,et al. 2006. Comment on “The Brain of LB1,Homo floresiensis.” Science 312:999.

179 Jacob T, Indriati E, Soejono RP, et al.2006. Pygmoid Australomelanesian Homo sapi-ens skeletal remains from Liang Bua, Flores:population affinities and pathological abnor-malities. Proc Natl Acad Sci USA103:13421–13426.

180 Weston EM, Lister AM. 2009. Insulardwarfism in hippos and a model for brain sizereduction in Homo floresiensis. Nature 459:85–88.

181 Schauber AD, Falk D. 2008. Proportionaldwarfism in foxes, mice, and humans: implica-tions for relative brain size in Homo floresien-sis. Am J Phys Anthropol 46(suppl):185–186.

182 Montgomery SH, Mundy NI. 2013. Parallelepisodes of phyletic dwarfism in callitrichidand cheirogaleid primates. J Evol Biol 26:810–819.

183 Morwood MJ, Jungers WL. 2009. Conclu-sions: implications of Liang Bua excavationsfor hominin evolution and biogeography.J Hum Evol 57:640–648.

184 Kaifu Y, Baba H, Sutikna T, et al. 2011.Craniofacial morphology of Homo floresiensis:Description, taxonomic affinities, and evolu-tionary implication. J Hum Evol 61:644–682.

185 Falk D, Hildebolt C, Smith K, et al. 2009.LB1’s virtual endocast, microcephaly, and hom-inin brain evolution. J Hum Evol 57:597–607.

186 Lordkipanidze D, Ponce de Leon MS,Margvelashvili A, et al. 2013. A complete skull

from Dmanisi, Georgia, and the evolutionarybiology of early Homo. Science 342:326–331.

187 Van der Geer A, Lyras G, De Vos J, et al.2010. Evolution of island mammals: adaptationand extinction of placental mammals onislands. Oxford: Wiley-Blackwell.

188 Holliday TW, Franciscus RG. 2009. Bodysize and its consequences: allometry and thelower limb length of Liang Bua 1 (Homo flore-siensis). J Hum Evol 57:223–228.

189 O’Connell JF, Allen J. 2015. The process,biotic impact, and global implications of thehuman colonization of Sahul about 47,000years ago. J Archaeol Sci 56:73–84.

190 Barker G, Barton H, Bird M, et al. 2007.The “human revolution” in lowland tropicalSoutheast Asia: the antiquity and behavior ofanatomically modern humans at Niah Cave(Sarawak, Boreno). J Hum Evol 52:243–261.

191 Storm P, Wood R, Stringer C. 2013. U-series and radiocarbon analyses of human andfaunal remains from Wajak, Indonesia. J HumEvol 64:356–365.

192 Pettit P, White M. 2012. The British Palae-olithic: human societies at the edge of the Pleis-tocene world. Abingdon, UK: Routledge.

193 Van Meerbeeck CJ, Renssen H, Roche DM.2009. How did Marine Isotope Stage 3 and LastGlacial Maximum climates differ? Perspectivesfrom equilibrium simulations. Clim Past 5:33–51.

194 Eriksson A, Betti L, Friend AD, et al. 2012.Late Pleistocene climate change and the globalexpansion of anatomically modern humans.Proc NatL Acad Sci USA 109:16089–16094.

195 Boivin N, Fuller DQ, Dennell R, et al. 2013.Human dispersal across diverse environmentsof Asia during the Upper Pleistocene. QuaternInt 300:32–47.

196 Parton A, Farrant AR, Leng MJ, et al.2013. An early MIS 3 pluvial phase in South-east Arabia: climatic and archaeological impli-cations. Q Int 300:62–74.

197 Krause J, Fu Q, Good J, et al. 2010. Thecomplete mitochondrial DNA genome of anunknown hominin from southern Siberia.Nature 464:894–897.

198 Meyers M, Fu Q, Aximu-Petri A, et al.2014. A mitochondrial genome sequence of ahominin from Sima de los Huesos. Nature 505:403–406.

199 Arsuaga JL, Martinez I, Arnold LJ. 2014.Neandertal roots: cranial and chronological evi-dence from Sima de los Huesos. Science 344:1358–1363.

200 Reich D, Green RE, Kircher M, et al. 2010.Genetic history of an archaic hominin groupfrom Denisova Cave in Siberia. Nature 468:1053–1060.

201 Pr€ufer K, Racimo F, Patterson N, et al.2014. The complete genome sequence of aNeanderthal from the Altai Mountains. Nature505:43–49.

202 Pennisi E. 2013. More genomes from Deni-sova Cave show mixing of early human groups.Science 340:799.

203 Huerta-S�anchez E, Jin X, Asan, et al. 2014.Altitude adaptation in Tibetans caused by intro-gression of Denisovan-like DNA. Nature 512:194–197.

204 Cooper A, Stringer CB. 2013. Did the Deni-sovans cross Wallace’s Line. Science 342:321–323.

205 Reich D, Patterson N, Kircher M, et al.2011. Denisova admixture and the first modernhuman dispersals into Southeast Asia and Oce-ania. Am J Hum Genet 89:516–528.

206 Weber M. 1902. Der australische archipelund die Geschichte seiner Tierwelt. Jena: Gus-tav Fisher Verlag.

207 Simpson GG. 1977. Too many lines; thelimits of the Oriental and Australian zoogeo-graphic regions. Proc Am Philos Soc 121:107–120.

208 Huxley TH. 1868. On the classification anddistribution of the Alectoromorphae and Heter-omorphae. Proc Zool Soc Lond 1868:296–319.

209 Audley-Charles MG, Hooder DA. 1973.Relation of Pleistocene migrations of pygmystegodonts to Island Arc tectonics in EasternIndonesia. Nature 241:197–198.

210 Hooijer DA. 1975. Quaternary mammalswest and east of Wallace’s Line. In: Bartstra G-J, Casparie WA, editors. Modern quaternaryresearch in Southeast Asia, vol. 1. Rotterdam:Balkema. p 37–46.

211 Murray-Wallace CV, Woodroffe CD. 2014.Quaternary sea-level changes: a global perspec-tive. Cambridge: Cambridge University Press.

212 Bird P. 2003. An updated digital model ofplate boundaries. Geochem Geophys Geosyst 4:1–52.

213 Zonneveld J-P, personal communication(e-mail, January 14, 2015).

214 Hall R. 2002. Cenozoic geological and platetectonic evolution of SE Asia and the SWPacific: computer-based reconstructions andanimations. J Asian Earth Sci 20:353–434.

215 Mac Arthur RH, Wilson EO. 1967. Thetheory of island biogeography. Princeton:Princeton University Press.

216 Van Valen L. 1973. Pattern and the bal-ance of nature. Evol Theor 1:31–49.

217 Heaney LR. 1978. Island area and bodysize of insular mammals: evidence from the tri-colored squirrel (Callosciurus prevosti) ofSoutheast Asia. Evolution 32:29–44.

218 Foster B. 1964. Evolution of mammals onislands. Nature 202:234–235.

219 Schwaner TD, Sarre SD. 1988. Body sizeof tiger snakes in Southern Australia with par-ticular reference to Notechis ater serventyi (Ela-pidae) on Chapelle Island. J Herpetol 22:24–33.

220 Grant PR. 1982. Variation in the size andshape of Darwin’s finch eggs. Auk 99:5–23.

221 Lomolino MV. 2005. Body size evolution ininsular vertebrates: generality of the islandrule. J Biogeogr 32:1683–1699.

222 Anderson RP, Handley CO. 2002. Dwarfismin insular sloths: biogeography, selection, andevolutionary rate. Evolution 56:1045–1058.

223 Mishra S, Gaillard C, Hertler C, et al. 2010.India and Java: contrasting records, intimateconnections. Q Int 223-224:265–270.

224 Puech PF. 1983. Tooth wear, diet, and theartifacts of Java Man. Curr Anthropol 24:381–382.

225 Pope G. 1984. The antiquity and paleoen-vironment of the Asian Hominidae. In: WhyteR, editor. The evolution of the East Asian envi-ronment. Hong Kong: University of HongKong. p 827–847.

226 Hutterer KL. 1977. Reinterpreting theSoutheast Asian Paleolithic. In: Allen J, GolsonJ, Jones R, editors. Sunda and Sahul: prehis-toric studies in Southeast Asia, Melanesia andAustralia. New York: Academic Press. p 31–71.

227 Bednarik RG. 2003. Seafaring in the Pleis-tocene. Cambridge Archaeol J 13:41–66.

228 van Es LJC. 1931. The age of Pithecanthro-pus. The Hague: Martinus Nijhoff.

229 von Koenigswald GHR. 1940. Preliminarynote on new remains of Pithecanthropus from


Central-Java. In: Proceedings of the Third Con-gress of Prehistorians of the Far East, 24th Jan-uary–30th January, 1938. Singapore: RafflesMuseum and Government Printing Office. p91–95.

230 von Koenigswald GHR. 1940. Neue Pithe-canthropus-Funde 1936–1938 Ein Beitrag zurKenntnis der Praehominiden. Dienst MijnbNed-Indie Wet Med 28:1–218.

231 Hooijer DA. 1956. The lower boundary ofthe Pleistocene in Java and the age of Pithecan-thropus. Quaternaria 3:5–50.

232 Hooijer DA. 1957. The correlations of fossilmammalian faunas and the Plio-Pleistoceneboundary in Java. Proc Kon Nederl AkadWetensch Amsterdam 60: 1–10.

233 Sartono S. 1961. Notes on a new find of aPithecanthropus mandible. Publikasi TeknikSeri Paleontologi 2: 1–51.

234 Sartono S. 1970. On the stratigraphic posi-tion of Pithecanthropus mandible C. Proceed-ings of the Institut Teknologi Bandung 4: 91–102.

235 Sartono S. 1975. Implications arising fromPithecanthropus VII. In: Tuttle RH, editor. Pale-oanthropology: morphology and paleoecology.The Hague: Mouton. p 327–360.

236 Ninkovich D, Burckle LH. 1978. Absoluteage of the base of the hominid-bearing beds inEastern Java. Nature 275:306–308.

237 Sartono S, Orchiston DW, Siesser WG,Djubiantono T. 1981. Upper Pliocene Sedi-ments in Sangiran, Central Java (Indonesia).Buletin Geologi Institut Teknologi Bandung 5:1–25.

238 Siesser WG, Orchiston DW. 1978. Micropa-leontological reassessment of the age of Pithe-canthropus Mandible C from Sangiran,Indonesia. In: Bartstra GJ, Casparie WA, edi-tors. Modern quaternary research in SoutheastAsia, vol. 4. Rotterdam: Balkema. p 25–30.

239 von Koenigswald GHR. 1968. Das absoluteAlter des Pithecanthropus erectus Dubois. In:Kurth G, editor. Evolution und Hominization.Stuttgart: Gustav Fischer. p. 195–203.

240 Nishimura S, Thio KH, Hefuwat F. 1980.Fission-track ages of the tuffs of the Pecanganand Kabuh formations, and the tektite at San-giran, Central Java. In: Nishimura S, editor.Physical Geology of Indonesian Island Arcs.Kyoto: Kyoto University. p. 72–80.

241 Orchiston DW, Siesser WG. 1982. Chrono-stratigraphy of the Plio-Pleistocene fossil homi-nids of Java. In: Bartstra GJ, CasparieWA, editors. Modern quaternary research inSoutheast Asia, vol. 7. Rotterdam: Balkema.p 131–149.

242 Hartung JB, Koeberl C. 1994. In search ofthe Australasian tektite source crater: the TonleSap hypothesis. Meteorit Planet Sci 29:411–416.

243 Gentner W, Storzer D, Wagner GA. 1969.New fission track ages of tektites and relatedglasses. Geochimica Cosmochim Acta 33:1075–1081.

244 Blum JD, Papanastassiou DA, Koeberl C,Wasserburg GJ. 1992. Neodymium and Stron-tium isotopic study of Australasian tektites:

new constraints on the provenance and age oftarget materials. Geochimica et CosmochimicaActa 56: 483–492.

245 Izett GA, Obradovich JD. 1992. Laser-fusion 40Ar/39Ar ages of Australasian tektites.Abstracts of the Lunar and Planetary ScienceConference 23: 593–594.

246 Guo S, Hao X, Chen B. 1996. Fission trackdating of Paleolithic site at Bose in Guangxi,south China. Acta Anthropol. Sinica 15: 347–350.

247 Hou Y, Potts R, Yuan B, et al. 2000. Mid-Pleistocene Acheulean-like stone technology ofthe Bose Basin, South China. Science 287:1622–1626.

248 Schneider DA, Kent DV, Mello GA. 1992. Adetailed chronology of the Australasian impactevent, the Brunhes-Matuyama geomagneticpolarity reversal, and global climate change.Earth Planet Sci Lett 111:395–405.

249 Lee M, Wei K. 2000. Australasian mictotek-tites in the South China Sea and the West Phil-ippine Sea: iImplications for age, size, andlocation of the impact crater. Meteorit PlanetSci 35: 1151–1155.

250 McCall J. 2001. Tektites in the geologicalrecord: showers of glass from the sky. London:Geological Society of London.

251 Langbroek M. 2015. Do tektites really datethe bifaces from Bose (Baise) Basin, Guangxi,southern China? J Hum Evol 80:175–178.

252 Wang W, Bae CJ. 2015. How old are theBose (Baise) Basin (Guangxi, southern China)bifaces? The Australasian tektites questionrevisited. J Hum Evol 80:171–174.

253 Pattan JN, Shyam Prasad M, BabuEVSSK. 2010. Correlation of the oldest TobaTuff to sediments in the central Indian OceanBasin. J Earth Syst Sci 119:531–539.

254 Dehn J, Farrel JW, Schmincke H-U. 1991.Neogene tephrochronology from site 758 onNinetyeast Ridge: Indonesian arc volcanism ofthe past 5 Ma. Proc Ocean Drilling Program SciResults 121:273–295.

255 Jones SC. 2007. The Toba supervolcaniceruption: Tephra-fall deposits in India andpaleoanthropological implications. In: PetragliaMD, Allchin B, editors. The evolution and his-tory of human populations in South Asia. NewYork: Springer. p 173–200.

256 Robock A, Ammann CM, Oman L, et al.2009. Did the Toba volcanic eruption of �74kBP produce widespread glaciation? J GeophysRes 114: D10107. doi:10.1029/2008JD011652

257 Ambrose SH. 1998. Late Pleistocenehuman population bottlenecks, volcanic winter,and differentiation of modern humans. J HumEvol 34:623–651.

258 Petraglia M, Korisettar R, Boivin N, et al.2007. Middle Paleolithic assemblages from theIndian subcontinent before and after the Tobasuper-eruption. Science 317:114–116.

259 Louys J. 2011. Mammal community struc-ture of Sundanese fossil assemblages from theLate Pleistocene, and a discussion on the eco-logical effects of the Toba eruption. Q Int 258:80–87.

260 Steiper ME. 2006. Population history, bio-geography, and taxonomy of orangutans(genus: Pongo) based on a population geneticmeta-analysis of multiple loci. J Hum Evol 50:509–522.

261 Wilting A, Sollmann R, Meijaard E, et al.2012. Mentawai’s endemic, relictual fauna: is itevidence for Pleistocene extinctions on Suma-tra? J Biogeogr 39:1608–1620.

262 Morwood MJ, Brown P, Jatmiko, et al.2006. Further evidence for small-bodied homi-nins from the Late Pleistocene of Flores, Indo-nesia. Nature 437:1012–1017.

263 Larson SG, Jungers W, Tocheri MW, et al.2009. Description of the upper limb skeleton ofHomo floresiensis. J Hum Evol 57:555–570.

264 Jungers WL, Larson SG, Harcourt-SmithW. 2009. Descriptions of the lower limb skele-ton of Homo floresiensis. J Hum Evol 57:538–554.

265 Stringer C. 2014. Small remains still posebig problems. Nature 514:427–429.

266 Tocheri M, Orr C, Larson S, et al. 2007.The primitive wrist of Homo floresiensis and itsimplications for hominin evolution. Science317:1743–1745.

267 Jungers WL, Harcourt-Smith WEH,Tocheri MW, et al. 2009. The foot of Homo flor-esiensis. Nature 459:81–84.

268 Weidenreich F. 1941. The extremity bonesof Sinanthropus pekinensis. Palaeontol Sinican.s. D, no. 5: 1–82.

269 Hershkovitz I, Kornreich L, Laron Z. 2007.Comparative skeletal features between Homofloresiensis and patients with primary growthhormone insensitivity (Laron syndrome). Am JPhys Anthropol 134:198–208.

270 Obendorf PJ, Oxnard CE, Kefford BJ.2008. Are the small human-like fossils found onFlores human endemic cretins? Proc R Soc B275:1287–1296.

271 Henneberg M, Eckhardt RB, ChavanavesS, et al. 2014. Evolved developmental homeo-stasis disturbed in LB1 from Flores, Indonesia,denotes Down syndrome and not diagnostictraits of the invalid species Homo floresiensis.Proc Natl Acad Sci USA 111:11967–11972.

272 Bowen DQ, Sykes GA. 1998. Correlation ofmarine events and glaciations on the northeastAtlantic margin. Philos Trans R Soc 318:619–635.

273 Willoughby PR. 2007. The evolution ofmodern humans in Africa: a comprehensiveguide. Lanham, MD: AltaMira Press.

274 Holloway RL. 1981. The Indonesian Homoerectus brain endocasts revisited. Am J PhysAnthropol 55:503–521.

275 Holloway RL. 1980. Indonesian “Solo”(Ngandong) endocranial reconstructions: somepreliminary observations and comparisons withNeandertal and Homo erectus groups. Am JPhys Anthropol 53:285–295.

VC 2015 Wiley Periodicals, Inc.


info:doi/10.1029/2008JD011652

early hominin biogeography in island southeast asia · early hominin biogeography in island...

Documents