You are on page 1of 7

NEW EVIDENCE ABOUT THE DATE OF COLONISATION OF NUKUORO ATOLL, A POLYNESIAN

OUTLIER IN THE EASTERN CAROLINE ISLANDS


Author(s): Janet Davidson
Source: The Journal of the Polynesian Society, Vol. 101, No. 3 (September 1992), pp. 293-298
Published by: The Polynesian Society
Stable URL: http://www.jstor.org/stable/20706460 .
Accessed: 25/06/2014 02:07

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

The Polynesian Society is collaborating with JSTOR to digitize, preserve and extend access to The Journal of
the Polynesian Society.

http://www.jstor.org

This content downloaded from 188.72.127.79 on Wed, 25 Jun 2014 02:07:49 AM


All use subject to JSTOR Terms and Conditions
SHORTER COMMUNICATION

NEWEVIDENCEABOUTTHEDATE OF COLONISATION
OF NUKUOROATOLL,A POLYNESIANOUTLIER INTHE
EASTERNCAROLINE ISLANDS
Janet Davidson
Museum ofNew Zealand, Wellington

The PolynesianOutliers are a numberof small communitieson the fringesof


Micronesia andMelanesia whose inhabitants speakPolynesian languages.They have
been variously regardedas relict colonies on the routeof Polynesianmigrations
eastwardsacross thePacific and theresultof backwashmigrationsand driftvoyages
afterthesettlement of Polynesia proper (Bayard 1976;Davidson 1970; 1974;Kirch
1984).Archaeological investigationshave shown thatboth of these theorieswere
undulysimplistic.At leastsomeof the islandsnow occupied byOutlier communities
have long and complex histories, reflectingthe varied patternsof colonisation,
settlement and culturalreplacementinmuch ofOceania (Davidson 1974;Kirch 1984).
The Polynesian languagesnow spokenon theOutliers and, insome cases, theorigin
traditionsof thepresent inhabitants,suggest thateffectivePolynesian colonisation
tookplace relativelyrecently;often longafterinitialhumancolonisationof the island
inquestion, as identified
by archaeology.However, archaeological identification of
Polynesian impact is often very difficultand usually needs to be supportedby
argumentsbased on linguisticor oral historyevidence. Consequently, theOutliers
continue to pose a particular challenge to archaeologists. In a paper on continuity,

changeand replacementon theOutliers,Kirch reviewedthearchaeological sequences


fornineof themand suggestedthatthetwonorthernmost,NukuoroandKapingamarangi,
"maywell exemplifytheeast-to-westdriftvoyagemodel ofOutlier settlement, with
relativelyshallow timedepth" (Kirch 1984:229). Two new radiocarbondates for
relativelyearly culturaldeposits on Nukuoro provide an opportunityto review this
assessment.

ARCHAEOLOGY ON NUKUORO
Nukuoro is the northernmostof theOutliers, and one of themost isolated.
Excavations on theatoll in 1965 (Davidson 1971) revealeddeeply stratifieddeposits
inmuch of thearea occupied by the singlemodem village on themain islet.The
depositswere rich in shell artefactsand faunal remains.Although therewere some
changes in stylesof fishhooks and other artefacts,thegeneral impressionwas of
continuityof occupation by a population alreadywell adapted toatoll life.Nothing

293

This content downloaded from 188.72.127.79 on Wed, 25 Jun 2014 02:07:49 AM


All use subject to JSTOR Terms and Conditions
294 JanetDavidson

was detected in the material recovered, and some indica


particularly "Polynesian"
tions of Micronesian affiliations or contacts were recognised.
A seriesof radiocarbondates suggested thatthebulk of thedeposits accumulated
relativelyrapidlyover a period spanningatmost about650 yearsfromthepresentday.
However, none of theearliestdepositswas dated, apartfroma sample fromnear the
base of theexcavation atNu-4, which initiallygave a resultof 20 B.C. ? 90 (Gak739).
A rerunof thissample gave a resultof less than 180yearsbeforepresent.Under these
circumstances, neither determination was as No further samples
regarded acceptable.
were available from this early deposit and consequently its age has remained
controversial.
The earliestdeposits locatedon Nukuoro containeddog remains,bothbones and
teeth.These depositswere found in threelocations in thecentralpartof thepresent
village and one in thenow uninhabitedinlandpartof the isletbehind thecentreof the
village.Dog remainsdisappear fromtheNukuoro sequence before thedepositionof
any of thedated layers(except thatunsatisfactorily dated byGak739). There was,
however,no stratigraphie evidenceof anyhiatus inoccupationbetweencultural layers
containingdog remainsand thosewithout.

THE RADIOCARBON SAMPLES AND THEIR CONTEXT


The development of theacceleratormass spectrometry method of radiocarbon
dating,capable of analysing much
very smaller samples,has provided theopportunity
fora freshlook at theage of theseearliestculturaldeposits on Nukuoro. No further
charcoal sampleswere available, butdog remainswere quite abundant.Collagen was
extracted from dog canine teeth from two locations and dated by the Nuclear Sciences

Group, DSIR Physical Sciences. The resultswere as follows.

NZA2I77 6,3C-15.41%o CRA 913?55B.P.


The calibratedage range,as calculated by the laboratory,isA.D. 1029 to 1243at the
95% confidence level. At the 68% confidence level, there are three intersects: A.D.
1047 to 1093 (23%), 1118 to 1144(13%), and 1153 to 1215 (32%), with amedian age
of A.D. 1135.

NZA2178 6,3C -12.47%o CRA 1237? 55 B.P.


The calibratedage rangeat 95% confidence isA.D. 680 to949. The 68% confidence
range isA.D. 724 to737 (5%) and 764 to889 (63%) with amedian age ofA.D. 815.
Sample NZA2177 isfromthesame contextinNu-4 as theunsatisfactory Gakushuin
results.Nu-4 is inthecentralpartof thevillage just inlandfromthemain path thatruns
parallel to theshore.The sample consistedof threeteethfrom layers3 and 4. Layer
3was a greyishwhite sandydeposit,capped by layer4, a thinlayerof black, charcoal
stained sand associated with two small firehearths,fromone ofwhich thecharcoal
sampleGak739 had been collected.This isnot theearliestevidence of humanactivity
atNu-4, for layer3 was underlainby layer2, an earlier layerof charcoal-stainedsand
on topof thebasal sterilewhite beach deposit.Layers 2,3 and 4 all contained sparse
artefactual and faunal material.

Sample NZA2178 consistedof two teethfromlevel7 inNu-6, whichwas situated

This content downloaded from 188.72.127.79 on Wed, 25 Jun 2014 02:07:49 AM


All use subject to JSTOR Terms and Conditions
New Evidence on theDate ofColonisation ofNukuoro Atoll 295

inthe low-lying,uninhabitedpartof the islet,about 60 metres inlandfromNu-4, and


about 15metres inlandfromthe inneredge of themain build-up of archaeological
deposits. The deposits in this locationwere poorly differentiated and excavation
proceeded in nine arbitrarylevels,numbered from the topdown. Artefactualand
faunalremainswere sparse.Dog remainswere recovered in level4 and levels6 to8,
suggestingthatthebottomhalfof thedeposit,at least,belonged to theearlypartof the
sequence on the atoll.
archaeological
The earlierof theseradiocarbonresultscomes close todoubling theknown spanof
humanoccupation on Nukuoro. Although some uncertainty must attend thecompa
rabilityof dog teethwith othermaterials as a medium fordating, there is no other
reason toquestion theresults.The new andmore reasonable resultfromNu-4 should
finallyenable theconflictingresultsfromGak739 tobe set aside.

DISCUSSION
Several interrelatedissues are raised by these results.There is thematter of
continuity of occupation throughouttheknown archaeological sequence; thereis the
matterof identifying Polynesian arrival; thereis thequestion of whether the result
fromNu-6 isan indicationof thedate of initialcolonisation.Underpinningall of these
isthevulnerabilityof atoll tandformstorapidgeomorphologicalchangeand theability
of archaeological deposits to survivesuch processes.
Atolls are subject toboth rapidaccretionand rapid lossof landas a resultof natural
processes. Human interventioninfluences islet-building,but the resultsof long
periodsof reclamationand erosionprotectioncan be swiftlyremovedby exceptional
storms.On Nukuoro islet,the lagoon foreshorein thevicinityof Nu-4 has probably
progradedbymore than40 metres since layers2 to4 were depositedand, inthe 1960s,
there was activeerosionof thelagoonforeshorefurther northinthevillage. Inprevious
, of theNukuoro sequence (Davidson 1971:18,29), itwas suggestedthat
interpretation
the basal layers of Nu-4 were the earliest encountered in the excavations, possibly
representing initial human colonisation of the atoll. It now appears, however, that the

deposits at Nu-6 are significantly earlier than those at Nu-4. It can be suggested,
therefore,thatprogradationof the foreshoreduringoccupation of the isletdid not
begin from a prehuman shoreline at Nu-4, but at some unknown point further inland
towards Nu-6.
The archaeological evidence fortheearlierpartof thearchaeological sequence on
Nukuoro is slightcomparedwith thesolid evidence forthemore recentpart,and it is
possible thatmost of theearlydeposits have been lost.The relativelysparse cultural
contentof theearly layersmay reflectoccasional activityby people whose actual
settlementwas elsewhere on the atoll, either on this islet or on another islet.
Investigationson theneighbouring Outlier of Kapingamarangi by Leach andWard
( 1981) suggesteda verysimilarpicture.They argued thatoccupationof theatollbegan
perhapsas much as 1,000yearsago butwas initiallycentredon one of the isletsother
thanTouhou (where the excavations took place). Touhou itselfhas been greatly
expanded and built up during thehuman occupation of theatoll,while deposits of
similarage on another isletor isletsappear tohave been completelystrippedaway.
This raises thepossibilitythattheearliestextantdepositsonmany atollsmay bemerely

This content downloaded from 188.72.127.79 on Wed, 25 Jun 2014 02:07:49 AM


All use subject to JSTOR Terms and Conditions
296 JanetDavidson

thefinal chapter inan extendedhistoryof repeatedbuild-upand stripping.


The findsassociated with the twoearlydated deposits on Nukuoro are relatively
few.However, thereisnothingto suggestdiscontinuity with the laterdeposits,other
thanthepresenceof dog remains,and theabsence of a fewartefacttypeswhich first
appear at various laterpoints in thesequence (Davidson 1971:28).
Layers 2 to 4 at Nu-4 yielded fragmentsof one-piece fishhooks and layer4
produced a fragmentary but recognisablebone lurepointwith a proximalprojection
and at least twoperforations. A similarbut largerfragment was found in the layer
above.Although lurepointswere notfound in laterprehistoricdeposits,similarpoints
and pearl-shelloccur inethnographiccollections fromNukuoro. Level
in turtle-shell
6 atNu-6, immediatelyabove thedated level,yielded a fishhookofType 1,a type
presentthroughoutthesequence. Ina recentstudyof fishingonKapingamarangi and
Nukuoro, thedeposits containingdog bones on Nukuoro were grouped as Period I.
Fish remains from thesedeposits showed an emphasis on trollingwhich declined
somewhat in laterperiods (Leach andDavidson 1988:15-17).
Nu-4 produced a complete but ungroundadze made from thedorsalmargin of
Tridacna maxima (previously incorrectly described as theventralmargin, see Kirch
and Yen 1982:210) and a fragmentof a ground adze. Level 6 at Nu-6 produced a
completeTridacna maxima dorsalmargin adze of the"normal variety" (Davidson
1971:56, 64-5) and a ground example without bevel. This adze sample is small;
however, the lackofTerebra orMitra adzes fromthesedated deposits and fromany
otherscontainingdog bones isan important point previouslyoverlooked.
Other itemsfromlayers2 to4 inNu-4 were twosmallwhite shelldisks and an un
finished small orange shell disk, a coral and a pumice abrader, a bevelled fragment of
a used bivalve (Asaphis violescens)
turtlecarapace and threeotherworked fragments,
and fragments of bone and pearl-shell.The only other finds from levels 6 and 7 inNu
6 were a used bivalve (Vasticardium elongatum) and some worked pearl- shell.
Bones of dog, rat, turtle, bird and fish were recovered from the early dated layers
in both locations.
The consistentpresenceof dog bones intheearliestdeposits,associatedwith only
limitedcultural remains,could suggest thepossibility that transienthuman visits
resulted in the establishmentof a feral dog population on Nukuoro which was
eliminatedbynewlyarrivedPolynesian colonistsat a laterdate.Regardless ofwhether
a dog population could surviveon an uninhabitedatoll, thereisgood reason toreject
thispossibility.Observation of a semi-feraldog population on the small islandof
Lakao in thetaumako Group near Santa Cruz duringfield work there in 1977-8
showed thesedogs tohave a veryhighdegree of toothwear as a resultof theirdiet.
There isno evidence of suchwear on thedog teethfromNukuoro.
The most reasonable interpretationof thepresentevidence of stratigraphy,chro
nologyand contentis that Nukuorowas continuouslyoccupied fromat leasttheninth
century A.D. There is no clear evidence of cultural replacement and nothing unam

biguously Polynesian rather thanMicronesian in the assemblage from the excavations.


The presence in the earliest levels of rat bones tentatively identified as a species found
inMicronesia but not Polynesia (Davidson 1971:91-2) is a strong indicationof
Micronesian influence.

This content downloaded from 188.72.127.79 on Wed, 25 Jun 2014 02:07:49 AM


All use subject to JSTOR Terms and Conditions
New Evidence on theDate ofColonisation ofNukuoro Atoll 297

The recentfindof two examples of theNukuoro Type I fishhook in theTemei


archaeological siteonVaitupu,Tuvalu, apparentlydated to the 11thcentury
A.D., has
been takenas theclearestpossible proof thatNukuoro was colonised fromTuvalu
(Takayama and Saito 1987:29). However, otherfishhooksfromTemei (Takayama
1987:24,25) do not show such close resemblancestoNukuoro. As previouslynoted
(Davidson 1971:40), thedistinctiveType I lineattachmentis seen on ethnographic
specimensof fishhooksfromPohnpei and fromotherPolynesianOutliers as well as
Tuvalu. The presenceof theType Ifishhook in level6 atNu-6 suggeststhatthistype
of hookwas presenton Nukuoro at leastas earlyas atTemei. The directionof contact
is as yet unclear.
Until quite recently,an initialcolonisationofNukuoro only600 or 1,000yearsago
seemedquite plausible.However, evidence has now accumulated suggestingthatthe
Marshall Islands, including the northernparts of the group, had certainlybeen
colonised 2,000 yearsago, and possibly considerablybefore that(Riley 1987; Shun
andAthens 1990;Streck 199p). It isclear thathabitable landexistedon these isletsby
3,000 yearsago (Dye 1987).Pohnpei andKosrae have also been inhabitedforat least
2,000 years,with a stillunfulfilledexpectation thattheywill be found tohave been
colonised early inthefirstmillenniumB.C. (Athens 1990;Ayres 1990).Under these
circumstances,it isunlikelythatNukuoro andKapingamarangi remainedundiscov
ered until late in thefirstmillenniumA.D. It is uncertain,however,whether any
archaeological evidence of theirinitialoccupations has survived.

CONCLUSION
The new radiocarbondates have extended the known occupation historyof
Nukuoro back to the8thor 9th centuryA.D. Inview of theage of settlement of some
Micronesian atolls,notablysomeof theMarshall Islands, itisunlikelythattheearliest
dated deposit represents initial colonisation of the atoll. However, it is not clear
whether earlier deposits have survived on Nukuoro. The vulnerabilityof atoll
landforms to severe damage and loss through natural processes is far greater than was

appreciatedat the timeof theexcavations on Nukuoro (cf.Davidson 1967 andDye


1987).
The pointatwhich Nukuoro became a PolynesianOutlier isno closer to recogni
tion.The older theoccupation sequence, the less likelyitis thatPolynesianswere the
only effectivecolonists.However, nothinghas occurred tochangemy earlierpessi
mism (Davidson 1970:72; 1971:104) about theabilityof archaeologists to identify
Polynesian arrivalonNukuoro, and I consider thesame toapply to thearchaeological
evidence fromotherOutliers thathave been investigatedarchaeologically.

ACKNOWLEDGEMENTS
The radiocarbondates were obtainedwith supportfrom theScience Research
DistributionCommitteeof theNew Zealand LotteryGrantsBoard. I am gratefultoJ.
McKee forhis interestindog teethas a datingmedium. Foss Leach andRogerGreen
commentedon a draftof thispaper.

This content downloaded from 188.72.127.79 on Wed, 25 Jun 2014 02:07:49 AM


All use subject to JSTOR Terms and Conditions
298 JanetDavidson

REFERENCES

Athens, J. S., 1990. Kosrae Pottery, Clay, and Early Settlement. Micronesica,

Supplement2:171-86.
Ayres,W. S., 1990.Pohnpei 'sPosition inEasternMicronesian Prehistory.
Micronesica,
Supplement2:187-212.
Bayard,D., 1976.The Cultural Relationships of thePolynesian Outliers. University
ofOtago Studies inPrehistoricAnthropology9.
M, 1967.Archaeology on Coral Atolls, inG. A. Highland,R.W. Force,
Davidson, J.
A. Howard,M. Kelly and Y. H. Sinoto (eds),Polynesian CultureHistory.Essays in
Honor ofKennethP. Emory.Honolulu: Bishop Museum Press, pp.363-75.
-1970. PolynesianOutliers and theProblem ofCulturalReplacement inSmall
Populations, inR. C. Green andM. Kelly (eds), Studies inOceanic CultureHistory.
Volume 1.PacificAnthropologicalRecords 11.Honolulu: DepartmentofAnthropol
ogy,Bishop Museum, pp. 61-72.
-1971. Archaeology on Nukuoro Atoll, a Polynesian Outlier in theEastern
Caroline Islands. Bulletin 9. Auckland: Auckland InstituteandMuseum.
-1974. CulturalReplacementon Small Islands:New Evidence fromPolynesian
Outliers. Mankind, 9:273-7.

Dye, T., 1987. Introduction,inT. Dye (ed.),Marshall Islands Archaeology. Pacific


AnthropologicalRecords 38. Honolulu: Departmentof Anthropology,Bishop Mu
seum, pp. 1-16.
Kirch, P. V., 1984.The PolynesianOutliers,Continuity,Change and Replacement.
Journal of Pacific History, 19:224-38.
-and D. E. Yen, 1982.Tikopia. The Prehistoryand Ecology of a Polynesian
Outlier. Bulletin 238. Honolulu: Bern iceP. Bishop Museum.
Leach, B. F. and J.M. Davidson, 1988.The Quest fortheRainbow Runner:Prehistoric
Fishing on Kapingamarangi and Nukuoro Atolls, Micronesia. Micronesica, 21:1 -22.
-and G. K. Ward, 1981. Archaeology on Kapingamarangi Atoll, a Polynesian
Outlier in theEastern Caroline Islands. Privatelypublished by B. F. Leach.
Riley, T. J., 1987. Report 2. Archaeological Survey and Testing,Majuro Atoll,
Marshall Islands, inT. Dye (ed.),Marshall Islands Archaeology. Pacific Anthropo
logicalRecords 38.Honolulu: DepartmentofAnthropology,BishopMuseum, pp. 169
270.
Shun,K. and J.S. Athens, 1990.Archaeological Investigationson Kwajalein Atoll,
Marshall Islands,Micronesia. Micronesica, Supplement2:231-40.
Streck,C. F., 1990. PrehistoricSettlement inEasternMicronesia: Archaeology on
Bikini Atoll, Republic of theMarshall Islands.Micronesica, Supplement2:247-60.
Takayama, J., 1987.The Western Origin forEarly Eastern Polynesian Fishhooks in
Light of theExcavation ofVaitupu, Tuvalu. Tezukayama UniversityReview, 57:16
46.
-and A. Saito, 1987.The Discovery of theDavidson Type laHooks on Vaitupu
Island, Tuvalu. Tezukayama University Review, 55:29-49.

This content downloaded from 188.72.127.79 on Wed, 25 Jun 2014 02:07:49 AM


All use subject to JSTOR Terms and Conditions

You might also like