ARCHAEOLOGY ON
KAPINGAMARANGI ATOLL
University of Otago
Studies in Prehistoric Anthropology.Volume
16
ARCHAEOLOGY ON KAPINGAMARANGI ATOLL
A POLYNESIAN OUTLIER IN THE
EASTERN CAROLINE ISLANDS
BY
Foss Leach
&
Graeme Ward
University of Otago, Studies in Prehistoric Anthropology, Volume 16
©
1981
Department of Anthropology, University of Otago
Copyright in all countries signatory
to the Berne Convention
(Cover design:
Canoe shed3
Werua isletj
i
Kapingamarangi)
Earlier volumes in the series Studies
in Prehistoric
Anthropology
are listed below, and may be obtained from: The Secretary,
Department of Anthropology, University of Otago, PO Box 56,
Dunedin, New Zealand. Those out of print are marked with an
asterisk(*). Prices quoted are in NZ$.
Volume 1*
B.F.Leach.1969. The Concept
Prehistoric
Studies.
of
Volume 2*
H.M.Leach.1969. Subsistence
New
Zealand.
Patterns
Volume 3*
J.I.Kennedy.1969.Settlement
Islands,
1772.
in
Volume 4*
D.T.Bayard.1971.Non Nok Tha.
Volume 5*
P.J.F.Coutts & M.Jurisich.1972. An
Archaeological
New
Zealand.
Survey of Ruapuke Island,
Volume 6*
M.Pietrusewsky. 1974.Non Nok Tha,
remains
from the 1966 excavations
N.E.Thailand.
Volume 7
C.F.W.Higham. 1975. Non Nok Tha:
Price: $3.50
Volume 8*
F.F.W.Higham, G.M.Mason, & S.J.E.Moore. 1976.
Upper Clutha
Valley.
Volume 9
D.T.Bayard. 1976. The Cultural
Relationships
Polynesian
Outliers.
Price: $3.50
Volume 10
P.J.Wilson. 1976. At our
Volume 11
A.Fox. 1978. Tiromoana
Pa, Te Awanga,
Excavations
1974-75.
Price: $4.50
Volume 12
P.Bellwood. 1 9 7 8 . A r c h a e o l o g i c a l Research
at Lake
Mangakaware,
Waikato,
1968-70.
Price: $4.50
Volume 13
D.T.Bayard. 1980. The Pa Mong Archaeological
1973-75.
Price: $5.00
Volume 14
C.F.W.Higham & Brian Vincent. 1980. Gabriels
an Archaeological
Survey.
Price: $4.00
Volume 15
Amphan Kijngam, Charles Higham, & Vvarrachai Wiriyaromp.
1980. Prehistoric
Settlement
Patterns
in
Northeast
Thailand.
Price: $8.00
Volume 16
Foss Leach and Graeme Ward. 1981. Archaeology
on
Raping amarangi
Atoll
- a Polynesian
Outlier
in the
Eastern
Caroline
Islands.
Price: $ 7.50
ii
Wit's
Similarity
the
in
in
South
The 1968
Prehistoric
East
Excavation.
the Human
at Non Nok
the
Bay of
Faunal
Skeletal
Tha,
Remains,
of
the
Beginning.Price:
Hawkes
$2.50
Bay
Survey,
Gully
-
CONTENTS
ACKNOWLEDGEMENTS
LIST OF FIGURES
LIST OF TABLES
INTRODUCTION
HISTORICAL SKETCH
ENVIRONMENTAL SETTING
General Comments
Climate and Ocean Currents
Pre-European Vegetation
Land Fauna and Avifauna
Marine Resources
Lithic Resources
Driftwood
Freshwater
ARCHAEOLOGICAL SITES ON KAPINGAMARANGI
The Site Survey
Culturally Modified Soils
Earthworks and Culturally redeposited
Coral Block Masonry
Traditional Sites
Historic Sites
Summary
THE EXCAVATIONS
Scope and Methods Adopted
The Excavation at Tiroki(KAl)
The Excavation at Muri-Harau (KA2)
iii
Page
The Excavation a t P u t a u (KA3)
42
The E x c a v a t i o n a t N g e i h o - H e r e u
(KA4)
RADIOCARBON DATING A N D T H E CHRONOLOGY OF SETTLEMENT
Level I: T r a n s i e n t Occupation
(71000 - 700 B P )
Level II:Formation of A r t i f i c i a l Islet(700-300BP)
Level III:Intensive Residential
Level IV:Modern H o r i z o n
46
50
53
53
Occupation(300-lOOBP)55
(100BP - P r e s e n t )
55
THE A N A L Y S I S OF M A T E R I A L S RECOVERED
Methods A d o p t e d
57
Fish
57
Shellfish
61
Crustaceans
63
Birds
64
Pig, Rat, Turtle and Human
65
Stone and Pumice
66
HUMAN SKELETAL MATERIAL
66
67
Shell Adzes
67
One-Piece Fish hooks
71
Trolling Lure shanks
73
Coconut Grater Heads
74
Pottery
75
Shell Arm Rings
76
Perforated Shells and Shell Beads
78
Drilled Sharks teeth
78
Sea Urchin Spine Files
79
Pumice Abraders
80
Worked Pearl Shell
80
IV
Page
Japanese and European Artefacts
81
KAPINGAMARANGI AND OTHER POLYNESIAN OUTLIERS
Linguistics
85
Drift Voyaging
86
Comparison with Nukuoro
88
CONCLUSIONS
Land History of the Atoll
91
Origin of the Founding Population
93
The Prehistoric Fishermen of Kapingamarangi
97
APPENDICES
1: D.T.Bayard. Marine Species in Kapingamarangi
orthography, with some proto-Polynesian
equivalents
101
2: Graeme Ward. Mollusc survey on Kapingamarangi
107
3: Foss Leach and Janet Davidson. The Analysis
of fish bone from Kapingamarangi
114
4: Graeme Mason. Kapingamarangi Crustacean Remains 123
5: Simon Holdaway. Analysis of Archaeological
shell from Kapingamarangi
125
6: W.R.Dickinson. Sand Temper in a sherd from
Kapingamarangi
134
7: C.A.Landis. Sand Temper Petrography of a
sherd and some Japanese tiles from
Kapingamarangi
135
8: Michiko Intoh and Foss Leach. Physical Analysis 136
of Kapingamarangi Pottery
9: Graeme Mason. Kapingamarangi Rocks
138
10:Phil Houghton. The Human Biology of
Kapingamarangi
14 2
145
v
ACKNOWLEDGEMENTS
The Kapingamarangi research was financed by the Historic
Preservation Office of the Trust Territory of the Pacific,
the University of Otago, and the Australian Institute of
Aboriginal Studies. This support is gratefully acknowledged.
We would like to thank Scott Russell and Ross Cordy of the
Historic Preservation Office in Saipan for their assistance in
a number of ways in organising the expedition, and also
Narsi Kostka and the Historic Preservation Committee in Ponape.
We are especially grateful to the Catholic Mission in
Ponape, and in particular to Father Cavanagh and Father Zewen
for their generous hospitality and interest in the work.
Thanks are also due to the Chief Magistrate and Council on
Kapingamarangi for their enthusiastic support of the research,
and to the people who assisted us on the island. Mike Lieber
helped us with a number of organisational matters, and this
assistance was very welcome.
The laboratory analysis of material recovered from
Kapingamarangi was assisted in various ways by Suzanne Moody;
specialist help was obtained from Ron Scarlett, Walter
Cernohorsky and Graeme Mason. Hank Jansen and Charlie McGill
provided much needed help with radiocarbon dating, and the
illustrations in this monograph are the work of Sandy McElrea,
Martin Fisher, and Richard Newell. To all these people, we
express our gratitude.
Janet Davidson provided a great deal of help and advice
on many matters, and we are very grateful to have been able to
draw on her experience with archaeological material from
Polynesian outliers. This manuscript benefited from her critical
eye.
Finally, we would like to thank Koro Monop and his
wife Temoe, our adopted parents on Kapingamarangi, who made
our visit to their island not only comfortable but memorable.
vi
LIST OF FIGURES
Page
1
Location Map of Kapingamarangi
5
2
Kapingamarangi Atoll, showing Islets
6
3
Kapingamarangi Wind Conditions through the year
7
4
The Islet of Pumatahati from a distance
16
5
Coral stone facing to taro swamp on Werua islet
17
6
Coral slab walls on Werua islet
17
7
Stone lining to fresh water well on Werua islet
19
8
The stone of Utamatua on Touhou islet
19
9
Stone fish trap between Tetau and Nikuhatu islets
20
10
Land building sea wall on Pepeio islet
21
11
Land building sea wall on Herekoro islet
21
12
Vertical Photograph of Touhou islet, 1942
23
13
A Japanese radio ship off Hare islet, 1942
24
14
Oblique aerial photograph of concrete structure
on Nunakita islet, 1943
25
15
The same building in 1980
25
16
Concrete house piles at Japanese installation on
Nunakita islet
26
17
Aerial photograph of Japanese weather station on
Nunakita islet, 1943
26
18
Concrete ablution structures at Japanese
installation on Nunakita islet
27
19
Propellors and engines from Japanese sea planes
on Hare islet
27
20
Map of northern half of Touhou islet showing
location of modern houses and excavation squares
30
21
Square layout and Section conventions
32
22
Vertical photograph of Werua islet in 1942
34
23
General view of excavation in progress at Tiroki
35
vii
Page
24
The stratigraphic sequence at Tiroki(KAl)
35
25
Section diagram of the excavation at Tiroki(KAl)
36
26
General view of the excavation at Muri-Karau(KA2)
39
27
Section diagram of the excavation at
Muri-Harau(KA2)
40
28
General view of the excavation at Putau(KA3)
43
29
The excavation at Putau(KA3) nearing completion
43
30
Section diagram of the excavation at Putau(KA3)
44
31
General view of the excavation at Ngeiho-Hereu(KA4)
47
32
The stratigraphic sequence at Ngeiho-Hereu(KA4)
47
S e c t i o n d i a g r a m of t h e e x c a v a t i o n at
48
33
Ngeiho-Hereu(KA4)
34
Profile of Touhou islet, showing the four
excavations
50
35
Correlation between the four sites excavated
52
36
The evolution of Touhou islet over the last 700 years 54
37
Tridacna shell adzes from Kapingamarangi
68
38
Surface collected Tridacna shell adzes
69
39
One-piece shell, bone and metal fish hooks
from Kapingamarangi
71
40
Shell trolling lure shanks from Kapingamarangi
73
41
Shell coconut grater heads from Kapingamarangi
74
42
Metal and ceramic artefacts from Kapingamarangi
76
43
Shell arm rings from Kapingamarangi
77
44
Files, shell beads and industrial pearl shell
from Kapingamarangi
79
45
A selection of European artefacts from the excavation 81
at Muri-Harau(KA2)
46
Glass beads from the excavation at Muri-Harau(KA2)
80
47
European artefacts recovered
at Muri-Harau(KA2)
82
viii
from the excavation
Page
48
Ceramic artefacts from the Japanese settlement
on Nunakita and Parakahi
83
49
The southern tip of Hare in 1942
91
LIST OF TABLES
1
Historical chronology of Kapingamarangi
4
2
Pre-European Vegetation on Kapingamarangi
9
3
Avifauna on Kapingamarangi
11
4
Radiocarbon Results for Kapingamarangi
51
5
Minimum numbers of fish organised into families
and chronological periods
58
6
Fish remains arranged in order of decreasing
abundance
59
7
Shellfish minimum numbers organised into families
and chronological periods
62
8
Bird Minimum numbers from the excavations
64
9
Minimum numbers of pig,turtle, rat and other
mammals from the excavations
65
10
C o m p a r i s o n of a d z e forms from N u k u o r o and
Kapingamarangi
69
11
Trends in the ratio of parrotfish to cod and
groper in Kapingamarangi
100
12
Basic results of fishbone analysis, Tiroki(KAl),
Minimum numbers
116
13
Basic results of fishbone analysis, Muri-Harau(KA2), 118
Minimum numbers
14
Basic results of fishbone analysis,Putau(KA3),
Minimum numbers
119
Basic results of fishbone analysis,Ngeiho-Hereu
(KA4), Minimum numbers
121
Crustacean remains from Kapingamarangi
124
. 15
16
ix
Page
17
The shellfish species identified from Kapingamarangi
arranged in taxonomic order
12 7
18
Minimum numbers of shellfish from Tiroki(KAl)
130
19
Minimum numbers of shellfish from Muri-Harau(KA2)
131
20
Minimum numbers of shellfish from Putau(KA3)
132
21
Minimum numbers of shellfish from Ngeiho-Hereu(KA4)
133
22
Physical analysis of pottery
136
23
Average stature figures for Kapingamarangi and
various other Oceanic groups
14 3
24
Blood group analysis of Kapingamarangi
144
-1INTRODUCTION
In 1978 the Chief Magistrate of Kapingamarangi wrote to
the Historic Preservation Office of the Trust Territory of the
Pacific drawing its attention to the fact that recent erosion
on Touhou islet had uncovered artefacts of kinds not known
to the present people, and stratified up to 4m below the
present ground surface. He requested that a suitable expert
should be sent to Kapingamarangi to investigate these finds,
and carry out a general programme of research into the ancient
history of the atoll, and make recommendations as to any future
work which should be done. The Historic Preservation Office
contracted the senior author of this report to organise an
expedition to the atoll for these purposes. Short notice prevented
a trip to the area before the southern University vacation of
1979-80.
A social anthropologist-Mike Lieber - has been working
on Kapingamarangi periodically for a number of years, and
his knowledge of the language and customs was seen as potentially
useful to the proposed archaeological research; consequently,
the Historic Prservation Office contracted him independently
to carry out research into traditional land use, and generally
assist the archaeologist with organisational matters.
Kapingamarangi is an isolated atoll, 750 km from the
port of Ponape, and shipping schedules are far from reliable.
However, a ship normally reaches the atoll once a month. The
research team was available from the period 16 November 1979
to February 16 in the hope that at least two full months
could be spent on the island. As it turned out, mechanical
problems with shipping resulted in delays which meant that
only one month of useful work could be carried out on the
island. Consequently, the excavation programme had to be
substantially limited in scope. In addition, Mike Lieber
had arranged to overlap his leave from University of Chicago
with the visit by the archaeologists. However, he was forced
to depart on the same ship which they arrived on. These
logistic problems were unfortunate, but other than the time
restriction did not adversely affect the research programme.
Following the fieldwork, three crates of archaeological
residues were shipped to Otago University for analysis. These
took seven months to arrive in New Zealand, and two were badly
damaged by forklifts somewhere in transit. Unfortunately some
material was lost, including a modern stone collection, and
some of the most important bones, not to mention items of
equipment and personal belongings. Lack of liability on the
part of shipping companies for such damage and loss is a little
known hazard which must be added to the list of difficulties
involved in archaeology in remote Pacific islands. An added
precaution was taken in the case of the radiocarbon samples,
which were air freighted to New Zealand at some expense.
These went missing at Nauru airport, and were the cause of
much anxiety until they were ultimately found.
Over the ensuing 12 months, the materials were analysed
in the Otago Archaeological Laboratories.
-2HISTORICAL SKETCH
Kapingamarangi has had a long history of irregular
contact with Europeans which is thought to have begun with
Fernao de Grijalvares in 1536. Emory, who has carried out
detailed analysis of historical information(1965:12ff) considers
that despite many sightings and possible stopovers, significant
contact with Europeans did not really begin until the last
two decades of the 19th century. The fleeting nature of these
early contacts is reflected in the numerous alternative names
and spellings for the island which were cited on various maps
and publications until relatively recently(see Baker,1951:14;
Emory, 1965:12-13;Schlaginhaufen,1929:219). Some of these are
listed below:
Kapingamarangi
Kapinmarangi
Baki ramarang
Kabeneylon
Kapenmailang
Makarama
Pikiram
Pigiram
Kapinga Malany
Pyghirap
Pyghiram
Puguirame
Pikiraf
Dos Pescadores
Tenuv
Constantine
Greenwich
Guru inichi
The first estimate of population size of about 150 people
was made in 1883(Eilers, 1934:6), but it is likely that the
pre-European population was somewhat larger than this. Emory
considers the climax population to have been about 450(1965:66).
About 1870, some castaways from the Marshall islands arrived on
Kapingamarangi, killed a large number of people, and dominated
the islanders until 1872 when the schooner Matautu took them back
to their homeland(Majuro and Ebon). An upper limit for a
traditional craft on such a journey would be about 3 0 people,
although there are reports of grossly overloaded Marshallese
outriggers with 40 to 50 aboard(Lewis, 1972:273). The number of
canoes which arrived on Kapingamarangi ranges in the telling from
three to eight, suggesting the arrival of between 90 and 240
people. It is hard to credit how such a number could subjugate
a population at least twice this size on their home territory.
This incident, along with traditional knowledge of Kapingamarangi
on Lukunor(in the Mortlock islands) and Nukuoro to the north
(Emory, 1965:52), and Ontong Java(Luangiua) to the south
(Schlaginhaufen, 1929:219), suggests that despite the isolated
position of this atoll, some contact with other areas might
have occurred from time to time in the prehistoric era. The
cultural significance of any such contact is a matter for
archaeological research to decide.
From 1536 to 18 99, Kapingamarangi was nominally a
Spanish territory, although their influence on the island
-3appears to have been minimal. The Germans purchased the Caroline
Islands from the Spanish in 18 9 9(presumably without consulting
the people on Kapingamarangi), and this began a period of
administration from Rabaul which lasted until the First World
War. During this time, the Japanese took over Micronesia,
and a weather station and seaplane base vere established on
the island during the Second World War. The Japanese base was
bombed in July 194 3. American administration of the island under
the aegis of the Trust Territory of the Pacific Islands began
in 1945, lasting until 1979. Kapingamarangi is now part of
the newly formed Federated States of Micronesia, and is
administered from Ponape to the north.
There have been a number of scientific surveys of
Kapingamarangi over the years, beginning with the German
South Seas Expedition of 1910, although Schlaginhaufen made
a brief visit there in 1908. Hambruch, who was writing up the
material from the Ponape district, died before this was completed,
and the Kapingamarangi section was eventually written up by
Eilers(1934). An interesting economic survey was carried out
in 1946(Gantt, 1946; Fosberg, 1946); and Elbert visited the
island in the same year and collected linguistic and folk-lore
material(Elbert, 1948;1949).Two major anthropological expeditions
to the island were organised by the Bishop Museum in 1947
and 1950, and the results were published in three detailed
bulletins(Buck, 1950; Miller, 1953; Emory, 1965). In 1954, a
further expedition took place, organised by the Pacific Science
Board, and this resulted in three further monographs describing
the geography, geology, and terrestrial bioecology of the
island(Niering, 1956;McKee,1956;Wiens,1956). Finally, from 1965
onwards, a social anthropologist has carried out a great deal
of fieldwork on the island, and on its satellite village
on Ponape,Porakiet(Lieber, 1968a;1968b;1970;1974;1977a;1977b;
Lieber and Dikepa, 1974) .
The combined effect of these researches has resulted in
an impressive body of literature on the island, its resources,
and its people. Until the Otago archaeology expedition, however,
nothing concrete was known of the prehistoric occupation
of the island. The various programmes of scientific research
on Kapingamarangi are listed in Table 1.
-4TABLE 1: HISTORICAL CHRONOLOGY OF KAPINGAMARANGI
FEDERATED STATES
OF MICRONESIA
1979
TRUST TERRITORY
OF PACIFIC ISLANDS
(American Admin)
1979-80 Otago University Archaeological
Expedition.Leach & Ward.4 0 days
1965-
Social Anthropological research
by Lieber
1954
Pacific Science Board Expedition
Wiens,Niering & McKee. 8 weeks
1950
Bishop Museum Expedition,II.
Emory & Miller. 5 months
1947
Bishop Museum Expedition I.
Buck,Elbert,Lanthrop & Emory.
3 months
1946
Visit by Elbert
1946
Economic Survey.Gantt & Fosberg
3 days
1943
Island bombed by US
1910
German South Seas Expedition
Hambruch,Kramer, etc. 6 days
1908
Visit by Schlaginhaufen.3 days
1899
Germans purchased Carolines
1870
Marshallese incident
1536
Discovery by
Fernao de Grijalvares
1945
JAPANESE PERIOD
1914
GERMAN PERIOD
1899
SPANISH PERIOD
1536
-5ENVIRONMENTAL SETTING
General Comments
Kapingamarangi is one of the most isolated areas of land
in the Pacific Ocean, and notoriously difficult to navigate to.
The only useful guideline for prehistoric navigators is the
'bird arc' of 7.5°. Commenting on a traditional voyage from
Pulusak(in the Western Carolines) to Kapingamarangi, a distance
of some 74 4 km, Lewis observed:
"It is tempting to surmise that, for such a difficult
passage(across variable currents) and devoid of any
additional reef or other safety 'screen', this 7.5°
expanded landfall arc might represent something like the
limit of navigational feasibility"(Lewis,1972:231).
Modern navigators to Kapingamarangi experience the same difficulty
in making a predictable landfall. Wiens(1956:21) recounts how
during their expedition to the island in 1954, the Captain of
10
3
9*
1
o cD
o
IE
o«
'\JOroluk
(^Truk
QPakin
AntoOPonape
Western Caroline atolls
o Mortlock Is,
£> Ngatik
^
^
oPingelap
O Kusaie
5• ° North
o Nukuoro
o kapingamarangi
-Equator '
CZZ? Admiralty Is,
New Ireland
- oTvkuu -
QNukumanU
• 5° SouthQ^pOntong Java
Bougainville
New Guinea
Figure 1: Location Map of Kapingamarangi. Some greater circle distances from
the atoll are: Nukuoro=215 km, Nuguria=511 km, Rabaul=651 km, Ponape=748 km.
Ocean currents in the vicinity of the atoll are thought to be generally
westward in character - experience shows them to be highly erratic.
-6-
t h e i r steamship s e t course from Nukuoro for a point 29 km to
the east of Kapingamarangi to compensate for the p r e v a i l i n g
westward d r i f t , but as a r e s u l t of a change in current d r i f t
to an e a s t e r l y d i r e c t i o n , they found themselves in e r r o r by
some 51 km to the e a s t ! This i s not an i s o l a t e d example, and
a t t e s t s the i s o l a t i o n of t h i s t i n y a t o l l from the r e s t of the
Pacific world(See Figure 1 ) .
Kapingamarangi i s about 100 km to the north of the equator,
and i s considerably c l o s e r to the major landmasses of Papua
New Guinea to the south than i t i s to i t s modern a d m i n i s t r a t i v e
centre of Ponape to the n o r t h . For t h i s reason, the Germans
considered i t appropriate during t h e i r administration to
include the i s l a n d in the d i s t r i c t of the Bismarck Archipelago,
r a t h e r than with the Eastern Caroline p a r t of t h e i r colony
(Schlaginhaufen,1929:220) . _
The a t o l l i s roughly egg shaped, with a maximum dimension
of about 12 km. The t o t a l area i s j u s t over 82 km2, of which
|
tf
V,
£•
a
/ / /
>
7
Jr
y ^^
jr
®
£i
o
46
c
^\\
~ X^.* *Xv/
J v J&.
»
a
o
m
d
79
79
0
73
46
f
^ s^ *X
X-^
J
Pungupungu
ft
73
}Wi^*
46
16
16
Ma tiro
fStJl'prtgt
Y^J^'Poe
/&]
Tetau
IS
X
37
I ^ P j Takairorigo
rt
o
I
16
H
J
%m
Cs
37
1
Spot soundings metre 5
North
S *
ue
X N
J6°33'ii9Bii
1"5
68
0
tf^s,^*
37
2&
69
<3
\
^I Fft \| Tout,ou
a
V^
i1
A
o
n
0
9
"
.^dr^'
»
*
\ ^"
5
0
•
1
ttfen
1
1
F i g u r e 2: Kapingamarangi a t o l l showing 34 i s l e t s on t h e c u r r e n t Hydrographic
C h a r t . Only 32 of t h e s e a r e now s e p a r a t e d by s e a . The southernmost land d i s t r i c t
on Hare was a s e p a r a t e i s l e t u n t i l r e c e n t l y , and both Herengaua and Herekoro
a r e now j o i n e d onto Hare(See a l s o F i g u r e 4 9 ) . Matukerekere was p r a c t i c a l l y
l e v e l l e d t o t h e sea i n a s e v e r e storm i n 1858(see Wiens,1956:19), b u t i s
growing a g a i n . The anchor symbol shows t h e l o c a t i o n of an anchorage n e a r Hare
where a J a p a n e s e schooner was sunk i n 1943(See a l s o F i g u r e 1 3 ) .
-7the reef and land ac count for a mere 2 0.4 km . The lagoon
reaches a depth of 8 0m in places. By comparison, the nearby
atoll of Nukuoro to the north is rather smaller in size(28.5 k m 2 ) ,
but with a slightly deeper lagoon(reaching 108m). Dry land on
Kapingamarangi total s about 112 hectares, and the 32 islets
(See Figure 2) range in size from 2.1 km in length(Hare) to only
3 0m long (Matukerekere). More than half the population of about
3 00 live on Touhou islet, which is about 300 x 150m in size,
covering 3.72 hectar es. Apart from about 20 people, the remainder
live on Werua islet, immediately to the north.
Climate and Ocean Currents
Precise information about the climate of this atoll is not
available, but several reconstructions have been made from patchy
observations over many years, and the comments here rely on
Wiens(1956:18-20,85ff) and Anon.(1944:3-4).The weather on
Kapingamarangi is controlled largely by the seasonal movement of
the doldrums belt, which lies immediately to the west and
south(Thomas,1965:30). In May and June the belt moves northward
across this area, and is accompanied by frequent heavy showers
and generally stormy weather. While the doldrums lie in the
north from July through September, generally fine weather
prevails. The doldrums move south again in October and November,
causing squally conditions interspersed with good weather.
From December through April the Northeast Trades predominate,
with good weather except for occasional tropical storms. Wind
frequency through the year is shown in Figure 3, and this
illustrates the influence of the Northeast Trade winds from
about November to April; the remainder of the year is
characterised by variable winds.
D
-CZ3
T\
D
Wind speed knots
Wind frequency percent
/J
40
60%
%)—c
calm 1
I
17
>27
Figure 3: Kapingamarangi wind conditions through the year(based on Anon.,1944:
4). A=December to February, B=March to May, C=June to August, and
D=September to November.
-8Rainfall on Kapingamarangi is highly -erratic from year to
year, largely because storms in the region are small in diameter,
and the number which actually pass over the island varies.
Average rainfall values ranging from 198 to 274 cm have been
published(Wiens,1956:18; Anon.1972:10). Temperature varies from
about 24° to 34°C, with a mean annual average of about 28°C.
Relative humidity varies from 57 to 90%, with an average of
about 78%.
Kapingamarangi lies in a region of the Pacific close to
the boundary of the south equatorial current(west to northwest
direction in this region), and the equatorial counter current
(easterly direction). Experiences by the Masters of various
ships over the years testify to very strong currents running
past the atoll - some claim these to be westward, others
eastward(Wiens,1956:21). It is a notoriously difficult place
to make a landfall at, which suggests an erratic character
in the ocean currents. The Pilot Charts for the region give
a predominantly westward current of 16 to 64 km per day(Anon.n.d.).
Waves and swell come from the northeast from December through
April, and largely from the southeast for the remainder of the
year.
Although hurricanes(typhoons) never occur on Kapingamarangi,
violent storms and tidal waves do. In addition, severe droughts
occur periodically.
Pre-European Vegetation
Trying to reconstruct in any detail the vegetation pattern
before the arrival of Europeans is very difficult. This has been
attempted by Niering (1956), Wiens(1956), and Emory(1965), but
it has to be admitted there there could be some important
errors of omission and commission. A summary of the position
is given in Table 2, from which it will be seen that the
flora was very restricted. Eighteen species of trees(of which
only about 10 are very common), four shrubs, and 18 herbs
are thought to have been available to the prehistoric inhabitants.
Of particular importance are those plants which may have
been brought by the founding population, or by later visitors
to Kapingamarangi after first settlement. Of these, particular
mention must be made of the taro. There are three species on
the island today - Cyvtosperma
ohamissonis,
the small swamp
taro, known as puraka, which is believed to have been introduced
from Nukuoro in Spanish times; Alocasia
maovorrhiza,
the large
dry land taro, again introduced recently along with the
name(ngaungau) from Nukuoro; and finally, Colooasia
esculenta,
the small common dry land taro, known as tava,
which is thought
to have been present in the prehistoric period(Emory,1965:7-8).
However, Thilenius believed differently, claiming that taro
was absent(Thilenius,1903:17). This issue could not be finally
resolved without positive evidence, such as from pollen spectra.
The fact that the name for taro is tava,
rather than the more
normal taro (by which it is known on Nukuoro), suggests that it
was present before contact with other communities in Micronesia,
such as people on Nukuoro.
-9TABLE 2: PRE-EUROPEAN VEGETATION ON KAPINGAMARANGI
This is largely derived from Emory(1965:8-9).Those
marked with an asterisk(*) are thought by Niering(1956)
to have been introduced by prehistoric Polynesians; the
remainder can be thought of as indigenous. The percentage
frequencies are taken from Niering(1956:Figures 28,29),
and are therefore based on modern vegetation density.
Those marked thus(t) are in the Niering list, but not
mentioned by Emory.
Trees
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
% Frequency
Cooos nuoifeva
Guettarda speciosa
Morinda oitrifolia
Premna obtusifolia
Pandanus tectorius
Barringtonia asiatioa
Calophyllum inophyllum
Hibiscus tiliaaeus
Cordia subcordata
Avtocarpus altilis
Messerschmidia argentea
Terminalia samoensis
Pisonia grandis
Thespesia populnea
Pandanus dubius
Ochrosia oppositifolia
Pongamia pinnata
Soulamea amara
100
97
97
97
97
63
51
48
45
42
39
27
15
12
6
6
3
3
*
t
t
97
36
27
6
t
*
*
*
*
*
t
*
Shrubs
1
2
3
4
Soaevola frutescens
Allophylus timorensis
Pipturus argenteus
Sophora tomentosa
Herbs
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Asplenium nidus
Lepturus repens
Stenotaphrum mioranthrum
Thuarea involuta
Cassytha filifovmis
Triumfetta prooumbens
Fleurya ruderalis
Ipomoea pes-aapvae
Ipomoea littoralis
Nephrolepis hirsutula
Wedelia biflora
Digitaria miorobachne
Euphorbia chamissonis
Ipomoea tuba
Jussiaea suffrutioosa (Ludwigia octovalvis)
Pteris tripartita
Colooasia esculenta
Psilotum nudum
54
36
36
33
18
18
12
12
12
9
9
6
6
6
6
6
3
3
t
t
t*
t
t*
t
*
-10Thus, we should think of the prehistoric cultivation
of taro on Kapingamarangi as making a striking contrast to the
extensive wet taro(puraka) cultivation of today, which covers
about 10 hectares(9.1% of the total land area,see Wiens,1956:36).
Whether the supposed prehistoric Coloaasia
esculenta
was
cultivated in swamp pits similar to those of today, or as an
essentially dry land form of horticulture, with perhaps some
kind of irrigation assistance, is not very easy to answer.
Breadfruit(Artocarpus altilis)
has the standard Polynesian
name(kuru),
rather than a term borrowed from Micronesia, and this
strongly suggests that it was either on the island when the
founding group reached Kapingamarangi, or it was brought by the
earliest settlers.
It will be noted in Table 2 that two herbs are thought
by Niering to have been introduced by man in the prehistoric
period, but are not considered by Emory. Digitaria
miorobaohne
is a grass which colonises open fern country, and
Ludwigia
oatovalvis
is a fairly common cultivation weed, especially
in Fiji and Samoa, and also common in swampy places(Parham,1972).
Why Niering believes them to have been introduced before the
arrival of Europeans is not clear.
Other plants of some economic significance today which
(arrowroot),
have been introduced recently are Taooa pinnatifida
Musa spp(banana), Carica papaya (papaya), and Curcuma
domestica
(turmeric). This latter introduction is especially interesting,
as it came from Nukuoro, where the plant possesses ritual
importance as it does elsewhere in Polynesia.
Land Fauna and Avifauna
Amongst the land animals of Kapingamarangi, those of
potential economic importance to prehistoric people would have
been land crabs such as Cardisoma
rotundum
and Coenobita
spp.,
and the coconut crab Birgus
latro.
Both skinks and geckos occur
on the island, but are unlikely to have been of any real value.
The Polynesian r a t ( E a t t u s exulans)
occurs, and there are a few
pigs and one dog on the island today. The question of when these
animals were first introduced will be discussed later.
Only a few birds occur on and about the island, but some
of these may have had prehistoric economic value. A list is
given in Table 3 of the main birds which are known on the
atoll. Of these, Aplonis
opaous is very common, and Gygis
alba
subsp., and Anous spp. moderately so. It is interesting that
large flocks of Fregata
minor subsp., as well as of Sula sula
rubripes
are on several of the islands. Emory notes that of the
common birds, only Sula sula rubripes
was ever eaten at all.
Observations during our expedition were to the contrary, and the
meat of the frigate bird, and others, was prized.
-11TABLE 3: AVIFAUNA ON KAPINGAMARANGI
Aplonis
opacus
Fregata
minor subsp.
Sula sula
rubvipes
Pluvialis
dominica
fulva
Anolls stolidus
pileatus
Anous tenuirostris
marcusi
Gygis alba subsp.
Phaethon
lepturus
dovotheae
Puffinus
Iherminieri
dichrous
Numenius phaeopus
variegatus
Crocethia
alba
Sterna
fuscata
oahuensis
Thalasseus
bergii
pelecanoides
Sula leuoogaster
plotus
Callooalia
inqwieta
inquieta
Eudynamis
taitensis
Gallus
gallus
Demigretta
sacra
sacra
Arenaria
interpres
interpres
Micronesian starling
Frigate bird
Red-footed booby
Pacific golden plover
Noddy tern
White-capped noddy tern
White tern
White tailed tropic
bird
Dusky shearwater
Whimbrel
Sanderling
Sooty tern
Crested tern
Brown booby
Carolines swiftlet
Long-tailed NZ Cuckoo
Domestic chicken
Reef heron
Turnstone
Marine Resources
The sea life around Kapingamarangi has unfortunately not
been studied in any detail by marine biologists. It represented
a major source of food for the prehistoric inhabitants, and
some background comments should therefore be made. In order to
identify the archaeological materials it was necessary to make
a comparative collection during the expedition, and although
the preparation of this should not be thought of as a systematic
survey, some familiarity was gained with marine life on the atoll.
Over 100 species of fish belonging to 37 families were
caught and prepared for osteological study. This work clearly
showed the lagoon to be abundant in a wide variety of fish, and
it is hardly necessary to go outside the lagoon in order to
catch them. The overall size of this resource makes it very
unlikely that it could have been depleted to any significant
degree by sustained exploitation by an atoll population, even
with the level of skills possessed by the Kapingamarangi
fishermen. Some species of fish, however, could dwindle in supply,
particularly those which are readily taken in the shallows
between islets - this includes members of the families
Muraenidae(moray eels), Diodontidae(porcupine fishes) and
Tetradontidae(puffer fishes). These factors are of importance
in assessing the significance of any trends in the nature of
prehistoric fish catches. It is interesting that lagoon fish
have apparently learned to keep at a greater distance from
divers with spears than previously; however, a good fisherman
finds no difficulty in bringing home a good catch after an
expedition on a fine day. The most reliable and productive method
of fishing today is certainly diving with locally made spear
guns. This is made possible by the recent appearance of
-12underwater goggles; and although some diving certainly took
place before their appearance, this method of fishing is bound
to have been far less productive in the past.
A list of the species caught is given in Appendix 1, along
with their local names. From this it will be seen that the most
abundant species are members of the cod and groper family
(Epiniphelidae), followed by unicornfish(Acanthuridae),
parrotfish(Scaridae), sea perches(Lutjanidae), and various kinds
of trevally(Carangidae). The most sought after fish, however,
are the pelagic game fishes, and in particular the tuna(Thunnidae
and Katsuwonidae), and to a lesser extent the wahoo(Aoanthoeybium
solandri)
. An especially important fish in chants is the rainbow
runner(Elegatis bipinnulatus),
which is also a surface feeding
fish, but a member of the Carangidae family. It is a common
European belief that all these fish can only be caught in very
deep water outside atoll reefs - this is simply untrue. Naturally,
they only occur some distance off land of high islands, but
they regularly enter lagoons of atolls, and can be readily
caught there. Fishing expeditions outside the reef take place
for more complex sociological reasons than supposed access
to wealthier fishing grounds.
Sharks are very common around the reef, both inside and
out. The people do not fear them, but swim freely amongst them,
and even play games with them. Although sharks are caught, they
are seldom eaten today, and one informant suggested their use
as fertiliser for taro beds after cooking them. There are two
species of crayfish at least, and these are reasonably common,
but not highly thought of as food. Marine crabs are also
numerous. Two species of marine turtle occur, but these are
fairly rare and small. The only browsing grounds for Zostera
sp.
sea grass are off Hare and Ringutoru, and these are only thin
and patchy. This also accounts for the general rarity of the
long toms(Belonidae). Cetaceans are occasionally stranded on the
reef, probably mostly killer whales{Orcinus oroa), and these
are butchered for oil and meat.
Shellfish resources are not as rich as they are on
many other islands in the Pacific, including atolls(See Appendix
2, and Table 17). The most significant species is undoubtedly
Tridaana
maxima, and specimens are quite abundant, especially
immediately to the seaward of the reef platform. Bippopus
hippopus
is the only other species of clam on Kapingamarangi, and although
this is less abundant, enormous specimens occur in deep water
in the lagoon. The main location of normal sized specimens is
in the sandy areas between islets, and in coral thickets in
fairly shallow waters of the lagoon near land. Of the two
species, only Tridaana
maxima is now considered worthy of
eating. Specimens of Turbo spp., especially Turbo
argyrostomus,
are common in some areas of the reef platform. There are of
course many other species of shellfish, but only a few can be
considered as of any real economic significance.
Of particular importance is the extreme rarity of specimens
spp.), or of black mussel(Atrina
vexillum)
of pearl shell {Pinotada
in the Kapingamarangi lagoon.
-13Buck comments on these as follows:
"The pearl oyster is scarce and the black mussel{woro)
can be obtained only in the main passage through the
reef. As they were sought for fishhooks, the flesh is
probably eaten"(Buck,1950:49).
Despite many days of searching for specimens by expert divers
all over the lagoon during our expedition, only one specimen
of each was located, and then only after the offer of a
substantial monetary reward. The pearl shell was located in only
moderately deep water on the southwest side of the lagoon, and
the black mussel was found in very shallow water off Hare.
There is a possibility that the black mussel was a recent plant
from another island(perhaps relating to the modern handicraft
industry), as it was found in a very odd ecological habitat
for this species. We can be quite certain that these two types
of shell do not constitute a significant resource, either for
food or for industrial material. The ecological reasons for this
dearth are not so easy to detect. Pearl shell is moderately
common on Nukuoro, an atoll which is fairly similar in general
structure. Of potential interest is the fact that specimens of
Spondylus
varians
are very common in the Kapingamarangi lagoon.
These are known locally as kainap,
and many are close to world
record size(290 mm being not uncommon ) . This shellfish, being
a sedentary creature in adulthood, would be a competitor with
Pinotada
spp. as far as habitat is concerned, though not
necessarily for food supply(Cernohorsky, 1981: pers.comm.)
Another potentially useful shellfish elsewhere in the
Pacific is Troehus
nilotiaus,
but no specimens of this were
recovered, either in the modern collections or archaeologically.
Further information on shellfish occurrence on Kapingamarangi
is given in Appendix 2 and Table 17.
Lithic Resources
Essentially all of the rock forming the islets on the atoll
is composed of calcareous skeletal debris, and rock fragments
cemented by calcium carbonate. It varies in texture and
degree of cementation; some being extremely friable, some well
cemented and dense(see McKee,1956). However, surprisingly
enough, other types of rocks do occur on Kapingamarangi beaches.
It is unlikely that these fragments belong to the rocky basement
supporting the coral atoll itself. Instead, it is believed
that these arrive primarily in the roots of floating trees,
and in other driftwood from elsewhere. Samples of rock from
driftwood trees were collected during the expedition, but
unfortunately all but two of the samples were lost in transit.
These are described in Appendix 9(specimens 79.64 and 79.294).
One of these rocks had an estimated weight of at least 150 kg
(specimen 79.64), although normally such rocks are more like
2 kg in size. It is quite certain that significant quantities
of rocky material have been arriving on Kapingamarangi from
high islands elsewhere in the Pacific for as long as the island
has been above the sea. These rocks are of considerable economic
-14importance as they are naturally favoured over limestone blocks
for cooking stones in ovans. In prehistoric times, these
fragments could also have been very useful for industrial
activities. The relative abundance of such rock was a considerable
surprise, and points to the need to be very careful about
leaping to conclusions of prehistoric trade and exchange based
on evidence of the foreign origin of rocks on an island. There
is a flourishing exchange network operating today, whereby
rocks for cooking ovens are imported in sacks on almost every
ship which visits the island. It would be interesting to study
the details of this network. Large flat rocks, used as anvils
for making Pandanus
sp. mats, are a feature of nearly every
household on Kapingamarangi, and these appear to have been
imported fairly recently, perhaps from Ponape. In the historic
era, stones have also arrived as ballast in ships now wrecked
on the island. For example, the well known Japanese ship
sunk off Hare(see Figure 13) contains large ballast boulders,
some of which are now in use on the island. Finally, pumice
occurs naturally on the beaches, and this has uses as abrasion
tools.
Driftwood
Very large logs and trees are regularly washed up, and
on atolls where large trees are either rare or absent, these can
be of great economic importance. Samples of a number of logs
were taken during the expedition, but these have yet to be
identified.
Freshwater
As mentioned above, Kapingamarangi has a highly variable
rainfall, and the amount of freshwater available is therefore
of critical importance. During the well known drought of
1916-1918, about 90 people died(Emory,1965:20; Wiens,1956:20).
This represented a large proportion of the population. Most
of the islets possess a lens of fresh or brackish water sitting
on top of the sea water which occurs in the base of the loose
sediments making up each islet. Wells are dug to reach this
water, and pits are dug to it for cultivating the recently
introduced swamp taro (Cyrtosperma
ohamissonis)
or puraka.
This water would normally be about 30cm above sea level, but can
reach 120 cm at times(see McKee,1956).
-15ARCHAEOLOGICAL SITES ON KAPINGAMARANGI
The Site Survey
The site survey was carried out with the aid of an
outrigger canoe with outboard motor over several days in the
company of a local informant who was knowledgeable about
land boundaries and island lore. Land boundary maps were
available from the report prepared by Lieber on traditional
land use(Lieber, 1979), and this proved most useful as the
assistant always knew exactly which part of an islet we were
at. This enabled detailed records to be kept on tape recordings,
selections from which were later transcribed into a running
diary.
Several problems became evident shortly after the survey
began, and the initial areas were covered a second time for the
sake of consistency. These problems can be enumerated as follows:
Firstly, informants kept referring to places where there
had been occupation during living memory but for which there
was a complete absence of normal archaeological evidence.
Kapingamarangi residents are meticulously tidy, discarding into
the sea all their rubbish. Consequently, there is an absence
of midden in occupied areas. Imported coral gravel
{kilikili)
can be identified as culturally deposited in two different ways.
The grain size can be a distinguishing characteristic against
the background of humified loam, or coarser coral rubble.
In addition, the kilikili
can be culturally modified by
subsequent activities, such as the incorporation of midden,
charcoal and ash; all of which again contrast with the general
background. Present residents try to keep kilikili
surfaces as
clear as possible, and cooking activities are carried out in an
area separate from the dwelling. This area can show
archaeologically as a small mound feature or as black soil
brought to the surface by land crabs. These types of evidence
certainly were found on the islets, but cultural soils
generally were poorly developed. This contrasts with nearby
islands such as Nukuoro and Ngatik, where cultural soils are
readily distinguished, and have a significant component of
midden as well as artefacts and industrial debris. The writers
saw more artefacts and industrial debris during two hours on
each of these islands, than during the entire site survey of
Kapingamarangi. This first problem of site identification
was confined to the low islets of the atoll; that is everywhere
except Touhou and Werua. Most surfaces are within one metre of
the high tide, so that exceptional tides and stormy weather will
certainly remove sediments and deposit rubble over site surfaces.
It is strongly suspected that some areas of low islets have
been physically stripped of occupation horizons by the sea
many times over the past millennium(see Figure 4 ) .
Secondly, the high islet of Touhou is at the opposite
end of the spectrum. Here there is so much culturally deposited
soil that the whole islet has to be considered as a single
archaeological site. The eastern side of Touhou has an exposed
deep section running along its entire seaward side. This shows
-16-
Figure 4: The islet of Pumatahati. This is a typical example of the low
lying character of islets on Kapingamarangi. Archaeological sediments of
any age do not occur on islets like this, which are periodically stripped
by storms and tidal waves.
very uniform modification of the imported soil throughout its
length, and could not be used to identify different activity
areas. With the exception of modern cooking places the same
continuity is evident on the surface of this islet too. The islet
of Werua lies between these two extremes of the low-lying
islets on the one hand and Touhou on the other.
Against this background it was clearly not possible to
employ the normal system of site numbering; rather, each islet
was annotated in the diary in terms of the types of sediments
found, and the character of any further archaeological evidence.
The archaeological sites found on Kapingamarangi can be placed
into several distinct categories as follows.
Culturally Modified Soils
Cultural soils were by far the most common feature seen.
Coral gravel has been brought onto the islets as kilikili
from
beaches and deposited around habitation areas, an activity that
can be observed today on Touhou and Werua. This is a very
common practice in Polynesia, and the normal process of
habitation alters the appearance of the gravel by the incorporation
of charcoal, ash, sparse food refuse, and other debris.
Consequently, kilikili
can be distinguished from naturally
deposited gravel by this cultural modification.
-17-
Figure 5: Coral stone facing to taro swamp on Werua islet
Figure 6: Coral slab walls on Werua islet
-18A few areas showed rather darker cultural soils which
are most likely to be associated with sustained cooking
activities rather than house foundations. This kind of soil,
however, was only rarely found during the survey. Genuine
midden areas, such as shell or fishbone dumps, are more or
less absent on Kapingamarangi. No doubt past communities did
what the contemporary inhabitants can be seen doing - throwing
their rubbish into the sea. Small quantities of midden do become
incorporated into and become covered by kilikili.
Again, this
process can be observed on the modern surface, and was also
evident in the excavations.
Earthworks and Culturally Redeposited Soils
Large quantities of natural sediments have been thrown
up during excavation of fields for taro cultivation and
redeposited in the form of high ridges surrounding the fields.
The cultivations are dug down to the surface of the freshwater
lens which occurs on top of the sea water which permeates the
substrate of these islets. The ridges can reach an exceptional
height of 6 m, such as those on Werua. They afford protection
from high seas. Lower ridges are inadequate in this role.
A severe storm, along the margins of a typhoon further to the
north, in December 1979 resulted in sea water getting into
many taro fields, thereby killing the crops.
Many of the taro field ridge sections were examined
without finding unambiguous evidence that the original
excavations had been made through an earlier cultural soil, or
buried natural horizon. This was a very surprising discovery,
and two possible interpretations can be suggested. Either, all
these excavations are amongst the earliest archaeological
features on Kapingamarangi, or periodic sediment stripping by
the sea means that whatever taro pits are present at any one
time are always the oldest archaeological features present,
and are continually re-cut into natural material.
Coral Block Masonry
Stone working, using coral blocks, occurs on Kapingamarangi
in several places, but is not as spectacular as that found on
some other Polynesian islands. The roads on Touhou and Werua
are outlined with coral slab uprights, as are some domestic
areas. Houses are only rarely outlined with actual curbing.
Some taro fields have dry stone facing to the interior,
such as on Ringutoru and Werua(see Figure 5 ) , and there is
also some dressed coral slab facing on the roadside on Werua
(Figure 6 ) . In addition, wells are invariably lined with coral
slabs(Figure 7 ) .
Land boundaries are designated by coral stone slabs which
might be free standing or incorporated into roadside curbing.
A traditionally important coral stone slab rests in an upright
position to the east of the present church on Touhou(see Figure 8 ) .
-19-
&j#
:»*»«<
Figure 7: Stone lining to freshwater well on Werua islet
Figure 8: The stone of Utamatua on Touhou islet.
-20-
Figure 9: Stone fish trap between Tetau and Nikuhatu islets
Free-standing sea walls have been built on both the
seaward and lagoon sides of Touhou. Much of the seaward wall
was destroyed by high seas recently, and was being reconstructed
during the period of fieldwork. Although communal activity was
involved on at least one occasion, the responsibility for
reconstruction was that of individual families whose land
holdings were being threatened, The discontinuous character of
the partially rebuilt walls had promoted the formation of
small beaches of coral gravel, which was the source of some
kilikili
being carried in baskets and bags to be deposited
around houses. Finer materials are obtained from sources on
other islets.
Fish ponds and fish traps were recorded during the
survey. One large pond, approximately 20 m in diameter, occurs
on the reef at the southern side, near the entrance of the atoll.
This was said to be used to store fish which are caught
nearby, but was in a state of disrepair when visited. Five fish
traps were said to have been present before the recent high
seas. The only one now extant occurs in the channel between
Tetau and Nikuhatu. This consists of a double line of boulders
arranged in a funnel shape in the shallows(Figure 9 ) . It is
used for channelling small fish(te tawin)
into the neck,
across which a fine net is placed.
Another important type of coral stone construction is
that of the land-building wall. This is built between two
blocks of land where there is a shallow channel, and it allows
the sea to throw up sediments behind it, thus creating new land.
-21-
Figure 10: Land building sea wall on Pepeio islet. Note the low-lying
nature of the land on the islet, and the effects of recent erosion.
Figure 11: An old land building sea wall on Herekoro islet which is now
well consolidated and colonised by trees and shrubs.
-22Several examples of this type of wall were found, notably those
at Pepeio, and at the southern end of Hare at Herekoro
and Herengaua(see Figures 10 and 11).
Traditional Sites
There are three notable traditional sites in Kapingamarangi.
The first is situated on the islet of Tipae, and is known as
the landing place of Utamatua, one of the founding ancestors
of the Kapingamarangi people. On the north side of Tipae is a
group of rough coral boulders forming a structure approximately
7m long and lm high. These are certainly culturally placed, and
although they now form an untidy pile, informants say that they
were carefully arranged until fairly recently, when they have
been disturbed by people seeking coconut crabs.
The second traditional place of interest is the tiny
sand patch called Matuke near Nunakita in the north. The
patch, which could once have been an islet, is now a rocky
outcrop only about 20 m in diameter. Utamatua is said to have
moved Matuke after being challenged by an earlier resident on
the atoll, and thus became the rightful owner of the land.
Utamatua may well have shrewdly awaited the onset of a storm
before taking up this challenge, because a high sea would
certainly lend some assistance in moving an islet in this
vicinity.
In the centre of Touhou is a large coral slab upright
known as satuhatu.
It is said to mark the burial place of
Utamatua(Figure 8 ) . This stone was earlier adjacent to the
cult house(Emory, 1965:206 ,-Figure 39), but this has been
replaced by the present day christian church.
Historic sites
Apart from recent evidence of occupation on the atoll
by the Kapingamarangi themselves, the most obvious historic
sites are those relating to the Pacific conflict of 1942-1945,
when the Japanese established military bases. There is also
evidence of the more recent US Air Force weather observatory,
and finally the remains of ill-fated Japanese fishing vessels
on the outer reef.
About 100m offshore from the Luawa land block
lagoon side of Hare lies the broken up remains of a
bomber(Emory, 1965:22) in about 10m of water. After
American divers exploded the bombs remaining in the
and this has further modified the wreck.
on the
B-17
the war,
aircraft
An area on Hare was the main gasoline and munition
storage dump for the Japanese seaplane base situated on the
lagoon side of Hare. Here still stands a square concrete water
tank. There is a series of bomb craters in the vicinity including
one massive one, apparantly representing a direct hit upon the
Figure 12: V e r t i c a l photo of Touhou i s l e t taken i n 1942. An American
I n t e l l i g e n c e r e p o r t i n d e n t i f i e d 70 b u i l d i n g s from t h i s photograph of a
p r e f a b r i c a t e d t y p e , s i m i l a r t o t h e Quonset h u t . F o r t u n a t e l y , i t was l a t e r
r e a l i s e d t h a t they were domestic houses of Kapingamarangi p e o p l e , o t h e r w i s e
they may have been bombed when t h e Japanese i n s t a l l a t i o n s were d e s t r o y e d
in 1943. A causeway i s under c o n s t r u c t i o n between Touhou and Werua t o t h e
l e f t of t h e photograph.
main munition dump. This bombing r a i d followed an a e r i a l
reconnaissance survey of Kapingamarangi in 1942(see Figure 12).
About 150m offshore from Hare(see Figure 2) i s the wreck
of a Japanese naval vessel which remains i n t a c t and in r e l a t i v e l y
good condition. This was recognised by US i n t e l l i g e n c e as a 38m
two masted a u x i l l i a r y schooner, with an extensive radio antenna
system. I t was sighted in the lagoon several times between
October 1942 and October 194 3. The ship appeared to be one of
a type of radio ship used by the Japanese to provide
communication in s t r a t e g i c areas and r e p o r t weather conditions
(Anon., 1944:1). The vessel r e s t s in about 12m of water. There
i s a double b a r r e l l e d a n t i - a i r c r a f t gun on a bow platform, and
there i s a marked f l a r e to the bows. Of p a r t i c u l a r importance
i s the fact t h a t the ship c a r r i e d a b a l l a s t of rocks. Several
large water-rounded rocks were recovered from the wreck by
diving, and c o n s t i t u t e the only l o c a l source of rock, other
than t h a t a r r i v i n g in the roots of d r i f t i n g t r e e s ( s e e Figure 13).
In an area between the US bomber and the Japanese vessel
j u s t mentioned i s a small Japanese m i l i t a r y s i n g l e - s e a t a i r c r a f t .
I t l i e s in about 4m of water, and i s now much scavenged by
divers.
-24At the northern end of Hare is a line of wrecked Japanese
sea planes resting in the sand of the lagoon, foreshore. There
are at least seven rotary engines, each with a triple-bladed
propellor(Figure 19). Immediately inland is a scatter of
tangled aluminium wreckage representing some of the fuselages
and wings of these aeroplanes. Aircraft identified by US
Intelligence are the Nakajima Type 97 fighter seaplanes, and
Kawanishi Type 97 four-engined flying boats which were used
in operations against Rabaul and Kavieng(Anon.,1944:1).
The islet of Nunakita has abundant evidence of the
Japanese occupation in the form of building foundations and
cultural debris such as porcelain scattered over the ground
surface(see Figures 14-18).
An American Air Force weather observation station was
situated until recently upon the southern islet of Pumatahati.
An extensive cleared area is being recolonised by low trees, but
Figure 13: A Japanese radio ship off Hare islet photographed in 1942. This
ship was later sunk, and its rock ballast is collected by divers for modern
use on the island.
-25-
Figure 14:Oblique aerial photograph of Japanese concrete two-story building
near a boat basin on Nunakita islet in September 1943. The purpose of the
structure puzzled American military intelligence(see Anon,1944:9).
Figure 15: The same building in 1980. It shows some damage from strafing with
plane machine guns, but is otherwise still intact. The bottom story is solid
except for a narrow vertical shaft. Emory states that this building housed
instruments in connection with the weather station(1965:21).
-26-
Figure 16: Concrete house piles at the Japanese installation on Nunakita islet
*
\
*
•
*
*
•••,
:
'
?•
•• •
%,
•
* *r
&
«r%
4
m**
.»
'v
-'
Figure 17: Aerial photograph of Japanese weather station on Nunakita islet,
September 1943 after bombing by the American military. In the foreground is a
15m observation tower, and other buildings are to the left(after Anon,1944:9),
-27-
r.lfr- ^ " " , - r f
Figure 18: Concrete ablution structures at the Japanese installation
on Nunakita islet. A water storage tank is in the foreground.
Figure 19: Propellors and engines from Japanese sea planes on Hare islet.
This base was destroyed in bombing raids on the island in September 1943
(see Anon,1944:1,10) .
-28rubbish dumps are still evident, along with three station
buildings in good condition.
Finally, there are two shipwrecks of modern Japanese
fishing vessels on the Kapingamarangi reef. One was driven on
to the north-western reef during the last few years. Everything
of value in the wreck has been salvaged by the people of
Kapingamarangi. The other vessel foundered on the.north-eastern
side of Nunakita about 20 years ago, and has been almost totally
destroyed by pounding seas. Remains of the engine and winches
can be seen on the reef in the surf, and the steering gear is on
the inner shore. The reef flat above the surge zone is scattered
with rusting iron fragments.
Summary
Kapingamarangi is not rich in deeply stratified
archaeological sites, and surface finds of artefacts are very
rare. During the site survey one piece of pottery, and five
shell adzes were collected. Considering the effort put into the
survey, this lack of return is remarkable; and along with the
absence of stratified midden debris is perhaps an indication
of the tidiness of the Kapingamarangi people. Today, roads are
swept, leaves are collected from around houses, and all debris
is committed to the sea.A significant buildup of cultural
debris is present only on Touhou, and for this reason, excavations
were concentrated there.
-29-
THE EXCAVATIONS
Scope and Methods Adopted
Within the available time of one month it was not possible
to explore more than a few lines of research; consequently,
strictly limited objectives were set and the excavation
strategy followed from these. First priority was given to
exploring the overall sequence of occupation of Kapingamarangi;
that is a diachronic approach to the atoll's prehistory. It was
necessary to abandon any hope of examining equally important
synchronic questions, such as those relating to settlement
pattern - this would have required large-scale area excavations,
which was simply not feasible.
It was hoped that in the course of an overall diachronic
study, sufficient midden debris would be found to assess the
character of Kapingamarangi prehistoric subsistence economics;
and that enough portable artefacts would be recovered both to
define the early material culture and to assess the possibility
of contact with other populations, such as those on Nukuoro
and islands further afield in Micronesia and Melanesia.
The site survey left no doubt that the place best suited
to documenting Kapingamarangi prehistory was Touhou islet. Here
there was every indication of a long sequence of intensive
habitation. Culturally introduced soils reached a depth of about
4m over much of the centre of the islet, and there was a good
chance of establishing a detailed occupation sequence.
Some general comments are appropriate about the specific
choice of location for the excavations. Factors influencing
choice of area to excavate vary depending on what the archaeologist
wants to learn by digging. For instance, investigation of the
settlement pattern in the period immediately before European
contact would involve one strategy, while an attempt to define
fishing equipment of the same period would involve another.
In most parts of Oceania, the choice of location for an excavation
to determine the latter would be closely related to whatever
could be learned from modern informants about the location of
traditional men's houses, canoe sheds, and so on. This is what
might be called the 'direct historical approach'. The relevance
of this strategy, however, to locating and excavating the
same type of cultural material from earlier periods becomes
increasingly questionable, the more removed the periods are
from the present day. This is especially so, on islets suspected
of having gone through significant changes in their
geomorphological character, during the supposed period of human
habitation. Archaeologists differ in their opinions as to just
where the cut-off point in time is. The present authors regar'd
an upper limit of about 300 years as the point at which one
must begin to employ quite different site location strategies.
As explained above, archaeological sites on Kapingamarangi are
not discrete entities in the way they frequently are in other
-30-
• KA3
C
0 £^
*KA1
KAi«
nS
D gnn
w1
JD ° a D' naif r
n1 ]
D nnn
50 Metres.
Figure 20: Map of the northern half of Touhou islet, showing the
location of modern houses and excavation squares. The dashed line
represents the profile given in Figure 34.
-31Pacific islands. This is especially so on Touhou, where the
entire islet is an archaeological site. Normally one has the
benefits of surface indications from sites which are spatially
separate from each other - that is, it is possible to gain
sample knowledge from several periods by examining the presentday surface. This is not generally possible on islets which are
artificially built up, and some other strategy must be employed.
Eroding sections of course are some help, but fail to give
any clues to what is in the centre of the islet - for this very
reason it is tempting to dig a hole in exactly this spot.
Geomorphological assumptions, such as the general process of
lagoonward shoreline progradation on atolls, provide one basis
for choice of location for examining different periods; but this
is not sufficient on its own. There is little point in digging
in a location hoping to find a 600 year old men's house, if
there is a good chance that the sea existed at this point
300 years ago. One should not underestimate the value of less
concrete factors, such as the 'pricking of one's thumbs', or
subtle variations in cliraatological aspect on an island, in
choosing where to dig. It has to be admitted that 'pot-luck'
may also be a significant factor on islets like Touhou.
Some consideration was given to all these factors in selecting
locations for excavation, without giving undue weight to any
one of them. The four areas excavated were deliberately chosen
to form an east-west transect of the islet. This general
approach had been adopted on Nukuoro by Davidson(1971:Figure 3 ) ,
and was a very successful way of revealing the overall history
of the island. It was hoped that a similar approach would work
for Kapingamarangi also(see Figure 20).
Since it was not possible to explore areal distribution
of evidence in the month available for the research, small
areas of 2 x 2m were laid out for excavation. It was hoped
that at least three and possibly four of these could be
completed in the time available.
Working in a confined space in coral g r a v e l ( k i l i k i l i ) s o i l s
is not without its hazards. For one thing, one cannot always
determine the precise origin of individual lenses of gravel,
which are sometimes re-deposited from earlier horizons following
the digging of postholes, ovens and the like. As other
archaeologists excavating in kilikili
soils have discovered,
there are numerous pseudomorphs and ghost structures which begin
and end with bewildering complexity. It is difficult enough to
trace these in plan during area excavation - it is well nigh
impossible in a 2 x 2m square. Some compromise was necessary,
and the unit level technique with 20cm spits was adopted as
most suited to the circumstances. Some particularly important
surfaces were exposed as they appeared and recorded separately;
but with these exceptions microstratigraphy was sacrificed.
All sections were plotted in detail, but some general
observations should be made here to explain how these are to
be interpreted.
-32Cultural stratigraphy to the archaeologist working on
coral islands tends to be a series of fine distinctions of
shades of grey. The layers are largely coral g r a v e l ( k i l i k i l i ) ,
some finer aggregates than others, some with lenses of charcoal,
some stained more than others, and so on. Most of these
distinctions have little or no recognised pedological meaning.
Often the lines observed in sections are undiscernible in plan
during excavation, and in any event these lines and lenses can
appear and disappear in an astonishing fashion which defies
simple stratigraphic interpretation. In this situation, it is
all too easy to lose sight of the wood for the trees. It is
therefore important not to become so baffled by the fine
stratigraphy during excavation that one fails to recognise
major changes which denote significant breaks in the function
of a site, or periods of surface stabilization, when these occur
Arriving at a reliable interpretation of the cultural meaning
of these major changes is another matter - it is obviously very
difficult in the context of a 2 x 2m square.
For these reasons, only changes which are thought to be
of some real significance are mentioned in the following
sections. By and large the stratigraphic changes recorded
Square Layout
Section Conventions
.. I^.U * ' ,!• ! I.jj.l... ..'."ilj'.'ltf' --.'•,
Humified Topsoil,
N
A
A
B
"^I'.y.'v.'r
•
. .':.-.'•.?:-"-''.f
•"•'••.•
t
Coral Gravel layer with
consolidated surface or
other significant change,
D
C
Sand and or gravel lens.
- 4 — 1 metre — I *
HWSf££TI£&&?&»
<?1&?SSS>*\
Ash lens
Charcoal Concentration.
Coral Boulder.
Figure 21: The four excavations were laid out in the manner shown above.
Sections are illustrated in Figures 25,27,30, and 33 using a series of
conventions described above.
-33probably represent periods of surface stabilization; this
explains the somewhat unusual simplified conventions adopted
for recording stratigraphy, and presenting them in section
diagrams(see Figure 21). In our opinion, such conventions are
entirely justified, and may be more meaningful in the sense
that they assist cross-comparison with other archaeological
sites.
When the overall stratigraphic sequence in each site
had been established with reasonable confidence, charcoal
samples for radiocarbon dating were taken from the sections.
The recovery technique involved trowelling out the spit matrix,
sieving it through one-quarter inch(6.35mm) screen, and hand
sorting on plastic sheets. All bone, shell, charcoal, noncoralline stone, and artefacts were recovered, and bagged
unwashed. The method of analysis of this debris is described
below.
In this manner, four excavations were carried out,
and approximately 54 cubic metres of material processed. In each
case, the excavation square was covered with a large parachute
to keep the exposed area dry.
The Excavation at Tiroki(KAl)
This area is in the vicinity of the highest part of
Touhou, about 70m inland(see Figure 20). Dark cultural material
was exposed at the surface, and it was an excellent location
in which to establish a deep stratigraphic history of the islet.
A single potsherd was recovered on the surface, the only one
found on Kapingamarangi(see Figure 42), and this was an
additional factor in choosing this spot for excavation.
According to Lieber(1979:122) informants on the island
disagree about the ancient activities on this plot of land.
However, one of the oldest men claimed that this plot was used
to dump trash from the northern side of the cult house area.
This was located where the present christian church is, and
is shown on Figure 20. It is not much further from the site of
the cult house to the sea - why it is thought that trash would
be dumped on this spot rather than thrown in the sea as it
normally would is not explained. There is also supposed to
have been a 'burial house' in the vicinity of Tiroki,
otherwise the plot was used mainly for growing coconut and
pandanus. The name Tivoki
literally means rubbish heap.
The plot of land is marked by several upright coral
slabs, which outline an approximate rectangle. The excavation
square was situated near the northeast corner, and it proved
to have about 4.5 m of deposit(see Figures 23 and 24). Using a
bench mark on the concrete wharf(+1.050mASL), a datum was set
up at Tiroki(+4.122mASL), and all relevant information was
referred to this point during the excavation. Full details of
this excavation and the others we^e recorded in the diary.
-34-
F i g u r e 2 2 : V e r i c a l photograph of Werua i s l e t t a k e n i n 1942. Note t h e
i n d i v i d u a l p l o t s of t a r o being c u l t i v a t e d i n t h e low swampy ground i n t h e
c e n t r e . This a r e a has been excavated t o t h e f r e s h w a t e r l e n s near sea l e v e l .
Immediately behind t h i s swampy ground i s a r i d g e of sediments which have
been thrown up t o a h e i g h t of 6m, and a f f o r d c o n s i d e r a b l e p r o t e c t i o n from
bad storms which a r e prone t o washing sea water i n t o t a r o p a t c h e s which can
t o t a l l y destroy the crop.
A s e r i e s of ten reasonably well defined l a y e r s were
encountered in the excavation, together with a number of
d i s t u r b a n c e s . These are shown in the section diagram, Figure 25.
The layers may be described as follows(NB: depth figures vary,
and are therefore only approximate).
Layer l(20am): humified top s o i l and c o r a l g r a v e l ,
with root h a i r s .
Layer 2(30am): loose c o r a l g r a v e l , with large q u a n t i t i e s
of fishbone. This m a t e r i a l i s c o n s i s t e n t with the
t r a d i t i o n a l information on the function of t h i s s i t e .
Layer 3 (30am): The surface of layer 3 was c l e a r l y
indicated by a compacted l i g h t grey s u r f a c e , marking a
period of surface stabilization The body of the layer
consisted of c o r a l gravel with some f i s h bone.
-35
Figure 23: The excavation at Tiroki(KAl) was on a low mound about 4m ASL, and
sited where a single potsherd was found on the surface(see Figure 42). The
entire depth of coral soil to below freshwater was culturally deposited. Only
the top 50cm has cultural material consistent with traditional knowledge of
the activities on this plot. In the background is a parachute, used to
cover the excavation during downpours which can arrive with little warning.
Figure 24: The stratigraphic sequence at Tiroki(KAl). The excavation went
down to 4.6m from the surface, and cultural material was found to 0.5mASL.
The curious varve-like features in Layers 5 and 7 are difficult to see from
this angle of view. A large oven feature is clear on the main West wall.
KA-1
North
-Ji..,-..1M.
; , . ' - '
. . . . " ' -
•' - .
'
South
L
. I I , . , 1.^111
'- '
- ' •
Excavation Limits
INZ 5135) 679BP±76
Lens of fishbone
18
19
<&
o-\
yy
20
ci?
££(*
Freshwater
level
( -0-55m ASL )
• 21
- 22
:
23
• 2L
-1 - «
Figure 25: Section diagram of the excavation at Tiroki(KAl). The large oven feature shown on the
West wall is clearly seen in Figure 24.
-37Layer 4(25 am): compacted brown sandy soil, cut into by
a number of features, indicating the use of the area for
cooking activities at this time.
.Layer
5(110 am): This was marked at the surface by a
zone of humification, representing an old stable surface,
but not a compacted one. The main body of the layer was
very unusual, consisting of at least 100 varve-like
silty structures, highly suggestive of water action.
In the body of the layer, a large wet patch was evident
showing that water still accumulates in the layer.
Layer 6(10 am):
humified loam.
dark brown to black charcoal stained
Layer
very similar in composition to Layer 5.
7(15
am):
Layer 8 (40am): This layer marks the beginning of white
coral sand and rubble which characterize the natural soils.
Four bands of sandy silt occurred in the layer, representing
points of surface stabilization. Cultural material rare.
Layer 9( 50 am): White angular coral rubble, and flake-like
sand. The layer is very natural looking; however, a lens
of culturally deposited fishbone occurred in the layer,
together with the distal end of a pearl shell lure
shank(see Figure 40).
Layer 10( > 110 am): flake-like white sand, and large
coral boulders. This is natural beach material, representing
the earliest phase of Touhou prehistory.
Features:
During the excavation, one massive oven structure
was uncovered. This shows on both the South and West
sections(see Figure 25). It had been used a number of
times, and had quite a complex structure to it. Two
independant charcoal samples from two different parts of
this oven complex gave similar ages of <250 years BP.
A large wet patch occurred in the body of Layer 5,
indicating that water either congregates or flows in this
spot for some unknown reason.
Two aspects of this stratigraphic sequence are particularly
important. Firstly, there is a clear distinction between
naturally and culturally deposited coral gravel at the junction
between Layers 7 and 8, occurring at 1.4mASL. However, cultural
material, including artefacts, midden, and charcoal, was found
to a depth of 0.5mASL in these naturally deposited soils. This
shows that people inhabited Kapingamarangi atoll when Touhou
islet was barely above the sea, and must have consisted at that
time of a low sandy flat, periodically washed by exceptionally
high tides, and certainly storms. A radiocarbon date of 679 + 78
BP, represents a time before the build-up of the cultural soils
which form Touhou islet today. Three dates in the bulk of this
artificially formed horizon indicate that the bulk of this
build-up, at this part of Touhou anyway, occurred within the
last 250 years.
-38Secondly, the soil complex represented by Layers 5 and 7
is somewhat enigmatic, but of potential importance. Water action
is strongly suggested by the varve-like appearance of the
section, and perhaps also by the large wet patch in the body
of the layer. One obvious possibility is mulching activity in
a taro patch. This has the effect of storing water, and could
perhaps form features of the kind observed. This interpretation
can only be advanced with reservation, however, for two reasons.
On the one hand it would imply quite a different economic
function of Touhou islet. As Emory has pointed out(1965:122),
Touhou was not a place for taro cultivation, in the historic era,
when the wet fields were not owned individually, but grouped
together and situated on other islets - largely Werua, Hare,
Ringutoru, and Torongahai. This same pattern exists on Nukuoro
also, but there is some suggestion there that some taro patches
until relatively recent times were individually sited in the
vicinity of family houses(Davidson,1981:pers.comm.). Such a
possibility could also be entertained for Kapingamarangi,
especially since it is believed that only Colocasia
esculenta
was cultivated in the pre-European period, and a change of
plot location is quite feasible. It is most interesting that
the height of this soil complex is considerably above the main
water table( 1.4 to 2.7mASL). The modern cultivation pattern
is to establish the taro field at the level of the GhybenHerzberg freshwater lens(see Figure 22). Of course this pattern
is of relevance to the cultivation of swamp taro,Cyrtosperma
ohamissonis,
and may not necessarily be most appropriate to
a largely different kind of crop. The freshwater lens can reach
a maximum height of about 1.7m above the salt water table(McKee,
1956:24), but would normally be only about 30cm above. On the
whole the origin of this layer is somewhat mysterious, but the
interpretation of a dry-land taro patch is cautiously put
forward. Some cultural material was recovered in the layer,
decreasingly common with depth. The subject of this layer is
returned to later.
The few artefacts recovered in the excavation are
described in detail below. At this point, mention can be made
of the find of a supposed distal fragment of a pearl-shell lure
shank in Layer 9. This is of some significance. It was intimately
associated with a radiocarbon sample which yielded an age of
679 + 78 BP. This establishes the existence of this unique
lure form to a period quite early in the Kapingamarangi
sequence. Only two shell adzes were found(see Figure 37), and
these are fairly late in age(Spits 3 and 8 ) . The faunal assemblage
is substantial, especially from the upper levels, entirely
vindicating the modern name of the site location. If the fish
bone is any indication, Spits 1 to 3 at this site represent
a more intensive dumping pattern than anywhere else
excavated(see Tables 12-15).
-39The E x c a v a t i o n a t
Muri-Harau(KA2)
The s e c o n d a r e a c h o s e n f o r e x c a v a t i o n was n e a r t h e
p r e s e n t l a g o o n s h o r e , i m m e d i a t e l y i n l a n d from s e v e r a l modern
c a n o e s h e d s , and n e a r t o an i m p o r t a n t m e n ' s h o u s e . The v i c i n i t y
of a c a n o e s h e d i s o f t e n assumed t o be an i d e a l p l a c e i n which
t o f i n d e v i d e n c e of f i s h hook m a n u f a c t u r e ( F u j i m u r a and A l k i r e ,
1 9 7 9 : 7 9 ) , and a t r a d i t i o n a l m e n ' s h o u s e i s c o n s i d e r e d a good
c h o i c e a l s o ( D a v i d s o n , 1 9 7 1 : 9 ) . Muri-Harau
l i t e r a l l y means
' b e h i n d t h e c a n o e h o u s e ' , and L i e b e r s u g g e s t s t h a t t h e r e was
one and p o s s i b l y two c a n o e h o u s e s h e r e a b o u t AD 1 8 0 0 ( L i e b e r ,
1 9 7 9 : i , 9 1 ) , as well as a ' b u r i a l h o u s e ' ( s e e a l s o Emory,1965:85).
A g e n e r a l p r o c e s s of s h o r e l i n e p r o g r a d a t i o n i n t h e d i r e c t i o n of
t h e l a g o o n i s now w i d e l y d o c u m e n t e d f o r i n h a b i t e d c o r a l
a t o l l s , and an e x c a v a t i o n s q u a r e i n t h e v i c i n i t y of M u r i - H a r a u
s h o u l d t h e r e f o r e n o t be t o o c l o s e t o t h e modern b e a c h , i f
d e p o s i t s of more t h a n a c e n t u r y o r s o a r e t o be f o u n d . A s p o t
25m back from t h e p r e s e n t s h o r e l i n e was c h o s e n ( s e e F i g u r e 2 0 ) .
The M u r i - H a r a u p l o t of l a n d i s q u i t e a c o m p l i c a t e d s h a p e
w i t h f i v e s i d e s . The e x c a v a t i o n was s i t u a t e d on t h e w e s t e r n
h a l f , a t 4.2m t o t h e w e s t of t h e p r e s e n t day roadway a r o u n d
t h e i s l e t . A datum a t +1.923mASL was e s t a b l i s h e d , and a l l
information r e l a t e d to t h i s p o i n t .
Figure 26: The e x c a v a t i o n a t Muri-Harau(KA2) i n p r o g r e s s . The c u l t u r a l
m a t e r i a l d e p o s i t e d h e r e reached a depth of only about 2m, and was l a r g e l y
r e c e n t i n o r i g i n ( u p t o 300 y e a r s ) . The s i t e i s immediately back from t h e
p r e s e n t lagoon s h o r e , and c l o s e t o a modern canoe shed. In t h e background,the
sediments a r e being s i e v e d and s o r t e d on p l a s t i c s h e e t s .
KA-2
meters ASL
North
East
2-
South
West
Sp.t
Layers
.•:•
Wff.
•
; '»
! : .*J :f^} ';'.-1
•
V>.
^:y•
;f;l^|-v;:---i-lfW'l^r^L:->:^
!;
i ,i,
*?.
tT-TT". ..: . . . , - , , . . , » . "
.-'IM I .:J1^J , ( ,MP'I..,.,... ? U J H'I^..IJI.J I
®
©
1 -
^
^
HIJ1.1^.^T^-UTJJ?JI•
'.wvyv.;-',;>•
.....i. r....' ,• I?.-!U.AI:I w. ^'.AL I U U / . J . - J I . u.»rJui»
^
M
^
^
.
^
L
L
^ V - ^ ^
U . < 1 r t^iMJj.J.'l J - . M . I r J
J J J . L-T.•
."JM . . , " U I J | , L W
;"^/,!,:;,liwwi.
* •• - • «•
»'—• : i:Ji'': :,:.,;!-,-.-,l.-'.!-:,'LT5 'I'jjkitLr•
»•
• • -i i-'-'-j"'-—•
-•• ' vnj..-y-iw<-''i'•-r.•
•»'•
•
r^aaiSr^s^S^p^^^ffiip-' ^ ^ ^ p > ^ ^ f e ^ ^ ^
' --•
.'.•
'.-•
<'•
•;'«
..
- j_
2^
I
ife.
O
I
^
©
J&&*?*"
««*£$
'
aw**
. .., :tf ; v >5»^i > - - -
Seawater level (-0.42 m ASL)
Excavation Limits
Figure 27: Section diagram for the excavation at Muri-Harau(KA2). European artefacts were recovered from
as deep as the stained zone in Layer 7. Some of these may date to the early European contact phase.
Note that gravel lenses occurring in Layer 7 slope down to the present lagoon shore. This shows the general
process of shoreline progradation over the past 300 years.
-41A series of seven moderately well defined layers were
discerned in the excavation(see Figures 26 and 27). These may
be described as follows:
Layer 1( 10 am):This was a complex layer consisting of
patchy humified coral gravel with several zones of
compaction.
Layer 2( 15 am): Loose layer of yellow sandy coral
gravel with irregular compaction.
Layer 3( 20 cm):
compaction.
Grey stained coral gravel with irregular
Layer
4(
Grey sandy-silty coral gravel.
Layer
5( 10 am):
Charcoal rich coral gravel.
Layer
6(
Grey sandy loam
15 am):
25 cm):
Layer 7(>100am):
Complex natural yellow beach sand layer.
The top 10-15cm was a zone of discoloration, either from
the leaching action of the upper cultural soil, or
initial occupation debris on the clean sand. Scattered
through the layer were a number of distinctive gravel
zones. These are tilted down in the direction of the
lagoon, and are progressively more horizontal later in
the site's history. This is clear evidence of progressive
shoreline progradation.
Features:
Several significant features were detected.
A hearth structure occurred in Layer 1, and another in
Layer 5. A large disturbance, possibly a huge post hole,
was also cut into Layer 5.
This site has two clearly defined phases, separated by the
junction between Layers 6 and 7. European material was found
right to this junction, and some of it may belong to the early
European contact period. A charcoal sample which gave the
single radiocarbon date for the site of <250 BP(see Figure 27)
was located in Layer 5. Some of the artefacts found are
illustrated in Figures 45-47; a few might be mentioned here.
An American 5C coin(1964), and a Japanese 10 sen piece(1942)
were recovered from Layer 2, and a Japanese 1 sen piece(1927)
was found in Layer 3. One pearl shell button(Figure 47, #F) was
found on the interface between Layers 6 and 7, and may well be
early European in age. A quantity of glass was found as deep
as Layer 5, and various glass beads were scattered through
Layers 3 to 5. Again, some of these items could be moderately
early, possibly back to Spanish times. Two metal fish hooks were
found from Layer 4(Figures 39,45). These are made in Polynesian
style, again suggesting some age.
A small amount of economic debris was found throughout the
site, extending below the main junction into Layer 7. This was
analysed by Layer rather than by spit as in the other excavations
-42The Excavation at Putau(KA3)
The third excavation was situated on a plot of land
known as Putau. Lieber suggests(1979:128), that about AD 1800
this area had a burial house as well as other houses for
residence, and a cooking area represented by a mound in the
North-east corner of the plot. Erosion processes have taken
away some of the scarp on the seaward side, and in 1979 some
human skeletal remains were exposed in the South-east corner
of the land. These were recovered by Lieber, and are described
in Appendix 10. The mound is a most interesting feature. The
surface layer is a fine black ash, and in this respect is very
similar to mounds on both Ngatik and Nukuoro. Only one such
mound now exists on Kapingamarangi, although it appears that
others may well have existed once. They are still fairly
numerous both on Ngatik and Nukuoro. They clearly represent an
area of concentrated cooking activity, and an excavation in
the vicinity could therefore be expected to yield valuable
economic information, and probably industrial material also.
The Putau area is a four sided parcel of land, and the
excavation was situated in the north-east corner at a position
2.6m back from the present scarp, and 7.8m from the north
boundary, directly on top of the mound. A datum was establishd
at a somewhat lower point than the top of the mound at 2.953mASL,
and all material is referred to this point. The location of
the site is shown on Figure 20, and general views of the
excavation are given in Figures 28 and 29.
A series of ten reasonably well defined layers were
observed(see Figure 30), and these may be described as follows:
Layer
1(
20 am):
Layer
2( 60 am):
humified top soil.
black sandy loam, with some coral rubble.
Layer 3( 70 cm): dark brown sandy loam. This is a complex
cultural soil, richer in ash and charcoal towards the
bottom.
Layer 4( 50 am): light brown coral gravel. This layer was
generally rather sterile, but rich in fish bone at the
top. There was loose coral rubble at the base.
Layer 5( 60 am): loose white coral gravel. This is a
complex cultural soil with various lenses and patches
of blacker material.
Layer 6( 20 am): dark grey sandy soil with patches of
white coral gravel and sand.
Layer
7( 20 am):
clean white sand.
Layer
8( 15 am):
black cultural soil
-43-
Figure 28: Laying out the excavation at Putau(KA3) on a low mound of black
soil about 3.5mASL. All of this deposit is artificial. To the right can be
seen the sea wall under reconstruction, following a massive storm in 1979.
Figure 29: The excavation at Putau(KA3) nearing completion at a depth of
4.1m. The consolidated basement of coral at this site was just below sea
level, and cultural material was recovered from its surface. The reflection
in the water is of the parachute covering the excavation.
KA-3
metres ASL
Spit
North
East
South
West
Layers
mA^'fW'k
*;-WMWWJ$^X»#vw
WT^T
j ^ . , :•• v..l^W^)?•p'
3-
•••'
•
.
•: '
• ( NZ 5i«) < 250 BP
©
<2>
—i
!
r-i"
•. - . . . .
.
- ' ••- .
'
, . . i....... i,.;,,.-,^,...,,/
,....,..•
. - ^ ' .••••
. '
.
i
rrra
2~
<tasfe
,-r_
3
:• ; -•••
—ftai.
^5^
0
1
-
(NZ51U)
• -•••i•
k
' TLI
-5
2 5 7 B P ±5 6
6
^
^.'-•
-',..,;......•
•
.•
.-•
™
•L'.ViJ'
: --.v_u.» ".•
i •. -
..
'
" - ™
.
• - ,'••. ' . . '
A -•- ' - '••
•
>-....
"•
^
^
^
' IJL i./ij^- >• - 1 -.yj J vtfit*- :• --.' T-
-•
7
8
•• -.'•
'• ' , ' \ 7 7
9
-^ssp^«*^^ c : ;.»<^
• 13
• •i •i'ii:-»i».',nvMl'.'.'.,-'j|iMj.i' ••|J—"j.*
yrt^^nVH^V/^J^^^^vL^J^i'w^LMMw-
%
—"
©
~
15
?kmwi&&wm® w&®@$^
j^^s^Jta^^w^p^sp^^^w5?^^^5££&££SS5£? 2&
16
1
'F^TTi•
.:^ > 1 .•• • -•.!.•
• , • • L ^ - , . . . . ^ r — ' • - •-• • • if•
u.-^WiJy.T*.'J.- . •• _
17
Remnant of old sea wall ?
©
riuLikHinng
18
.
.
.
.
,._•
.
-
Compacted Coral Basement'
1
9
|y
21
\ 22
Figure 30: Section diagram of the Putau(KA3) excavation. A rich cultural layer was present below sea level.
-45Layer
9(
60 cm):
Layer
10(20
am):
white coarse angular coral gravel and sand.
grey sandy coral gravel - cultural soil.
Features:
Few features were noted in this excavation.
However, in the top part of Layer 9 some flat coral slabs
were present face down, and these may have been the remnants
of an early retaining wall of some kind.
This is a complex archaeological site. Its function as a cooking
area does not extend very far back in time, probably only to
the base of Layer 3. Below this point, the cultural layers are
far less rich in charcoal, and the surface of Layer 5 particularly
is marked by the appearance of very natural looking material. The
bulk of the deposit below the surface of Layer 4 is interpreted
as part of the artificial build-up of Touhou islet. Two points
are notable about this material. Firstly,it does not consist of
carefully selected kilikili
coral gravel, but is rather a patchwork
of sand and coral rubble and gravel. This was probably obtained
from the nearby reef flat area. The material should therefore be
interpreted as deliberate build-up or fill material, rather than
fine quality kilikili,
such as is deposited in and around house
floors. Secondly, there are a number of layers of richer cultural
material which occur during this period of build-up; and this
shows that the process was a progressive one rather than a single
concentrated operation.
It is most interesting that cultural material occurs right
on the indurated coral basement of Touhou islet. As with the
evidence from Tiroki(KAl), this suggests that people were
present on Kapingamarangi at a time when Touhou was virtually
awash with the sea. Clearly, these people must have inhabited
some other islet at this time, and only temporarily visited
Touhou. The fluctuating sea level is shown on Figure 30, and it
will be noticed that the lowest cultural layer occurs below sea
level. Tidal fluctuations on Kapingamarangi vary from about 60
to 120cm through the lunar cycle(Wiens,1956:22;Niering,1956:
Figure 9 ) , and the coral basement at Putau is about -0.60mASL.
Thus, Layer 10 is always slightly under sea water. For this
reason, the material in Layer 10 is not necessarily in a position
of primary deposition, but is interpreted as part of the general
tide wash from activities being carried out on Touhou when the
islet existed as a low sandy flat, barely above the sea.
Four of the five radiocarbon samples give a clear picture
of the chronology of events at Putau. These are shown on Figure 30,
and are further discussed below. It may be noted here that the
upper horizon(above Layer 4 ) , consisting of the cooking mound
function, dates back to about 250 to 300 years only. The period
of artificial build-up of Touhou precedes this by a further 400
years to about 700BP, and the lowest cultural horizon(below
Layer 8 ) , is somewhat earlier again. The charcoal sample from
Layer 10, unfortunately does not provide a reliable estimate
of the age of this layer.The reason for the divergent result
is obscure, but being immersed in sea water,with circulating
modern carbon, for up to a millennium must have something to
do with it.
-46Artefacts were again rare in this excavation, but some
are important enough to mention at this point. A further distal
fragment of the characteristic Kapingamarangi pearl shell lure
shank was recovered from the gravel lens shown towards the
base of Layer 4 in Figure 30. This item was unfortunately
amongst the material lost in transit with the freight containers,
and is therefore not illustrated. The closely associated
radiocarbon sample dated to 5 35 + 5 7 BP, provides another good
estimate of the antiquity of this lure form. Four Tridaana
sp.
adzes(Figure 37), and a Cassis
sp. coconut grater head(Figure 41)
were found in the excavation. Of particular note too are two
triangular-sectioned Tvidaona
sp. shell arm rings, found in
Layer 10(see Figure 43).
A large quantity of midden was recovered in the excavation,
and this is described below.
The Excavation at Ngeiho-Hereu(KA4)
The fourth excavation was carried out more or less in the
centre of Touhou(see Figure 20), in a parcel of land known
as Ngeiho-Hereu,
meaning north of the cult house. This is the
largest and most important compound on Touhou. Lieber's
records show the existence here at about AD 1800 of residential
houses, burial houses, sacred houses connected with the cult
house, and a large cooking area(Lieber, 1979:95-96). In the
late 19th century, a men's house was also established here.
The cooking area was once marked by a mound, comparable to that
at Putau, but this was levelled during the Japanese period.
There was a reasonable chance that an early men's house or
canoe shed might be located deeply buried in this vicinity,
considering the likelihood that the early lagoon shore once
existed this far inland. The three excavations carried out so
far had failed to yield much evidence of fish hook manufacture it was hoped that this final excavation would.
An excavation square was laid out as near as could be
judged to where the cooking mound once existed, 6.7m north of
the curbing of the present pathway in the centre of Ngeiho-Hereu.
A datum was established at 2.572mASL, and all material related to
this point.
A series of six well defined deep cultural layers were
found in the excavation, and these are described as follows(see
Figures 31-33).
Layer
1(
35 am):
Humified black sandy coral gravel
Layer 2( 70 am): Grey ash. This layer evidently represents
the remaining part of the cooking area function of this
site.
Layer 3( 80 am): Black coral gravel layer. This material
is definitely kilikili
gravel, showing the site to have
had a residential function at this time.
-47-
Figure 31: The area at Ngeiho-Hereu(KA4) cleared of surface rubble immediately
before excavation. A mound feature relating to cooking activities in this
vicinity was levelled during the period of Japanese occupation of the island.
Figure 32: The stratigraphic sequence at Ngeiho-Hereu(KA4). The excavation
reached 3.7m in depth. Towards the base can be seen irregular coral boulders
which may have been part of an early wall structure. Cultural material in
this site extended to below sea level, and into the fresh water lens.
KA-4
North
East
South
West
metres ASL
Figure 33: The section diagram of the excavation at Ngeiho-Hereu(KA4). The boulders in Layer 5 are
probably part of an early wall structure of some kind. Layers 3 and 4 are residential kiZ'l'ki'l-i,
while
Layers 1 and 2 relate to the traditional cooking activities of this site.
-49Layer 4( 70 cm): Yellow-grey coral gravel layer. Again,
the coral gravel is kilikili
in origin.
Layer 5( 60 am): Yellow sandy coral gravel with numerous
coral boulders, and some lenses of cultural soil.
Layer
6(>70
cm):
Yellow sand with natural coral gravel.
Features:
The site had quite a few lenses of ash, charcoal,
and gravel in various layers; but other than this it was
a relatively featureless site. A large charcoal lens
occurred in Layer 3. The large numbers of coral boulders
in Layer 5 may have originally constituted some kind of
wall structure, but this is not certain.
Two radiocarbon samples provide age estimates of 354 + 56 BP
for Layer 3, and 548 + 78 BP for Layer 5. The kilikili
gravel in
Layers 3 and 4 suggests that the area was primarily residential
at this time, and was followed by the cooking activities of
the upper 2 50 years or so. The earliest cultural remains in
Layers 5 and 6 in the site were deposited in what is essentially
natural material. Layer 5 may represent deliberate building up
of the land above sea level, using nearby rocky sandy fill.
The distinctive tilted strandlines found in the basal sediments
at Muri-Harau(KA2) did not occur in this site; this suggests
that the lagoon shoreline has been further to the west than
this part of Ngeiho-Hereu for at least 600 years.
Only a few artefacts were recovered in the excavation.
Worthy of note at this point are a Conus sp. shell arm ring(see
Figure 43), and a shell coconut grater head(see Figure 41), both
from Layer 3; three one-piece fish hooks(Figure 39) from
Layer 4, and four adze fragments from Layers 3 and 4(Figure 37).
The presence of fish hooks in Layer 4 may indicate the existence
of a men's house in this vicinity about 400 to 500 years ago.
A considerable quantity of midden material was recovered,
and this is described below.
-50RADIOCARBON DATING AND THE CHRONOLOGY OF SETTLEMENT
Touhou islet, small though it is, has an area of about
3.72 hectares. The four excavations carried out constitute a
sample of about 0.04% of this area. On Nukuoro, Davidson
worried that her 4 0 m 2 of excavations was not a very large
sample(Davidson,1971:27); the 16 m 2 explored on Kapingamarangi
may seem cause for even less security, when it comes to overall
interpretation. However, the excavation and analysis of 54 cubic
metres of archaeological debris represents a sizeable quantity
of material to work with, and it is therefore believed that a
reasonably reliable picture can be put together of the overall
chronology of settlement on this islet, and perhaps for
Kapingamarangi as a whole.
In Figure 34 a profile of Touhou is illustrated. This
shows that the land is somewhat higher on the seaward side,
sloping away to the lagoon shore. The excavations went to
well below sea level in each case, but only at Putau(KA3)
was an indurated coral basement struck. This shows that Touhou
is really only a pile of gravel perched precariously on the
inner edge of a reef of coral rock. Cultural material was
present almost to the base of each excavation, and this
demonstrates that almost the entire volume of Touhou islet
above sea level(estimated at about 96,000 m 3 ) has been
redeposited by human agency. The extent of this build-up qualifies
Touhou for the appellation 'artificial island' in the same
sense as those recorded in the Solomon Islands at Taumako,
and along the coast of Malaita. This particular character of
Touhou is a striking contrast to the remaining low-lying islets
in the Kapingamarangi atoll chain. The fact that Touhou continues
to exist in its present elevated form is due quite simply to
the continued human maintenance of a coral block wall along
the seaward margin of the islet. In order to establish the
sequence of events involved in building up Touhou to this
present form, the deposits in the excavations must be correlated.
To assist this process, a series of twelve charcoal samples
were radiocarbon dated. The position of all samples is shown
on Figures 25,27,30, and 33. The samples of charcoal-rich soils
East
77777
Coral basement
Figure 34: Profile of Touhou islet showing the four excavations, and
following the transect illustrated on Figure 20 Vertical scale exaggeration
five times horizontal. Each excavation went to below sea level.
-51-
were of course contaminated with coral, and pre-treatment was
therefore necessary. The samples were sieved through one-quarter
inch (6.35mm) mesh, and all larger coral discarded. The charcoal
rich residue was treated with 2% HsPOi* until all CO2 evolution
from the coral had ceased. The sludge was then washed, neutralised,
combusted wet, and counted for 1000 minutes. The results are
given in Table 4 below:
TABLE 4: RADIOCARBON RESULTS FOR KAPINGAMARANGI
Provenance
T=5568
T=5730
T=5730*
6 13 C
NZ5135
KA1 Layer 9
679 + 78
699 + 81
695 + 65
-24.2
NZ5136
KA1 Layer 3/4 <250
<250
<250
-25.7
NZ5137
KA1 Layer 3/4 <250
<250
<250
-23.2
NZ5138t
KA1 Layer 6
<250
<250
<250
-24.2
NZ51395
KA2 Layer 5
<250
<250
<250
-23.3
NZ5140
KA3 Layer 8
714 + 58
734 + 59
710 + 59
-23.1
NZ5141
KA3 Layer 4
535 + 57
551 + 59
560 + 64
-23.8
NZ5142
KA3 Layer 10
558 + 57
574 + 59
591 + 59
-25.3
NZ5143
KA3 Layer 2
NZ5144
KA3 Layer 3
257 + 56
264 + 58
362 + 58
-25.2
NZ5145
KA4 Layer 3
354 + 56
364 + 58
430 + 58
-25.0
NZ5146
KA4 Layer 5
548 + 78
564 + 81
573 + 86
-25.5
<250
<250
<250
-24.4
NB: * = Corrected for Secular effect
Nominal ages are as follows:
NZ5138t
223 + 56
229 + 58
303 + 68
NZ5139§
117 + 56'
120 + 57
152 + 89
Although these nominal ages cannot be taken too seriously,
they are entirely reasonable estimates for their contexts.
The only additional absolute dating evidence comes from the
coins in KA2. Coins dated to 1964 and 1942 were found in Layer 2;
thus 1964 provides a terminus -post quern for Layer 2. In addition,
the 1927 coin found in Layer 3, provides a terminus
-post quern
for Layer 3.
a.
m
1/1
KA-1
L_
lKA-3
0
Level m
2
1
__.^--
2
100
3
Level m
\
3
^--—'
4
<250BP
5
6
7
<250BP
8
<250BP
257BP+56
wn'
--*,
* -^
8
9
10
^^
2
N,
'
\
3
9
4
10
5
<250BP
11
535BP+57
—
•
7
"
13
8
12
14
9
13
15
14
16
7%BP±58-
700
-
—
•— " "
17
"
*
18
V
19
18
20
19
21
20
"21
-10
679BP+78
11
558BP+57
(suspect)
22
\
\
\
3-
KA-2
V
\
\
\
35^'BP±56
\
\
\
<250BP
12
\
••" " "
\
\
11
16
\
6
12
15
nnn
1
x>^
5
7
Level I
KA-4
4
S
Level n
1/1
1
\
13
\
\
\
s
.14
7
15
8
16
5U8BP±78
9
17
\
\
18
\
19
Figure 35: The correlation between the sediments in the four sites excavated.
10
\
\
11
— OmASL
-53With one exception, these results form a very satifactory
series. The result for NZ5142 is not considered reliable.
In view of the importance of the cultural material below sea
level in Putau(KA3) Layer 10, it was regarded as essential to
attempt to date a charcoal sample from this provenance, despite
the fact that any material here had been washed continuously
with sea water for the duration of Touhou prehistory. Most
archaeologists would agree that this is tempting fate, and the
result is not satisfactory. In our opinion, the sample has been
contaminated with modern carbon, although the precise mechanism
is obscure.
Taken together, the dates show that the Touhou archaeological
deposits have formed fairly rapidly, and that the prehistoric
sequence for Touhou,at least, is not much older than about
700 years. Taking into account the sequence of cultural and
natural stratigraphy outlined above, the following reconstruction
is offered(see Figures 35 and 36).
Level I: Transient Occupation ( ? 1000 to 700 BP )
This period is represented by sediments and associated
material found at two sites only:
Tiroki Layer 10
(KA1 Spits 17-22)
Putau Layers 9 and 10
(KA3 Spits 17-21)
In both cases, the sediment matrix is essentially natural, but
cultural material is mixed into the sediments. Putau at this
time was only 0.25mASL, but Tiroki was slightly higher at
0.5 0mASL. There would not have been a natural freshwater lens
on Touhou at this time, and it is most unlikely that the islet
was permanently occupied. However, there certainly were people
on Kapingamarangi atoll somewhere, because they left scattered
cultural material in these early sediments on Touhou. These
people must have had a village or villages on one or more
of the other islets. It is quite likely that Touhou functioned
as a small outpost, in much the same way as Matukerekere does
today. That is, it is periodically visited, temporary shelters
are made, and it is used as a base for fishing expeditions.
The lower age of this period can only be guessed at.
Level II; Formation of Artificial Islet ( 700 to 300 BP)
This period is represented by sediments in three sites:
Tiroki Layers 8 and 9
(KA1 Spits 13-16)
Putau Layers 4 to 8
(KA3 Spits 8 -16)
Ngeiho-Hereu Layers 3 to 6
(KA4 Spits 7 -19)
This building up process was obviously quite complicated. At
Ngeiho-Hereu, for example, Layers 5 and 6 represent redeposited
natural sediments, while Layers 3 and 4 are clearly imported
-54-
Figure 36: An idealised picture of the evolution of Touhou islet over
the last 700 years. Vertical scale exaggeration five times horizontal.
Note that the lens of fresh water could not have always been present on
the islet, but would have begun to form as the islet grew in size.
-55residential kilikili
coral gravel. At Putau, most of this
fill material is redeposited natural, but lenses of rich
cultural soil occur as well. Finally, at Tiroki, the fill is
distinctly natural in appearance, but again lenses of cultural
material are found. Touhou islet by the end of this period had
been built up to a height of 2.OmASL next to the open sea(KA3),
1.8mASL on the lagoonside(KA4), and there was a hollow in the
middle of the islet(see Figure 36), where the height was only
about 1.4mASL(KAl).
Level III: Intensive Residential Occupation ( 300 to 100 BP)
This period is represented by material in the four sites
excavated:
Tiroki Layers 3 to 7
(KA1 Spits 3-12)
Muri-Harau below crust Layer 7 (KA2 Spits 7-11)
Putau lower half of Layer 2
(KA3 Spits 3- 7)
Ngeiho-Hereu Layer 2 and
top of Layer 3
(KA4 Spits 3- 6)
These sediments are the product of residential activities
consistent with those reconstructed for Touhou at about AD 1800
by Lieber(1979). In particular, the deposits at Putau arid
Ngeiho-Hereu throughout this period are ash and charcoal rich
gravels indicative of specialised cooking areas. On the
lagoonside at Muri-Harau, the sediments are natural, and
show that the high tide shoreline a century ago was still
further inland than it is today. The sediments at Tiroki for
the early part of this period(Layers 5 and 7 especially) are
distinct, and somewhat enigmatic. One possibility is that they
represent the accumulated material in a waterlogged low lying
area in the centre of Touhou where cultivation of taro took
place. The deposits are well above the Ghyben-Herzberg fresh
water lens, and this area could not be thought of as an example
of the kind of taro patch which is found on Kapingamarangi
today. These are essentially designed for the wet cultivation
of Cyrtosperma
ahamissonis
. A possibility which has to be
suggested is that this area of Touhou was used for the cultivation
of Coloaasia
esculenta,
the only taro form believed to have
been on Kapingamarangi before European contact. This taro could
have been grown in a dry-land garden, with mulching to improve
water retention properties of the soil.
Level IV: Modern Horizon
( 100 BP to Present)
The upper six layers of Muri-Harau, and the crust of
Layer 7(KA2 Spits 1-7) contain abundant European cultural
items such as those made from glass, iron, and brass; and it
appears that the bulk of this site should be placed in a
relatively recent time bracket. For convenience, the upper two
spits of the other three sites are also placed in this period.
-56This designation is somewhat arbitrary, especially in the case
of Ngeiho-Hereu, where some of the upper deposits have been
removed in the last 30 years; but it serves to separate these
superficial deposits from earlier ones for the purposes of
midden analysis.
-57THE ANALYSIS OF MATERIAL RECOVERED
Methods Adopted
When the archaeological material arrived at Otago, it
was all rebagged using a uniform system of labelling such as
KAl/A/Spit 6. Paper bags were used as they permit slow drying
which greatly increases the strength of archaeological materials.
At subsequent stages of rebagging, after drying was complete,
self-sealing plastic bags were used for permanent storage. All
identifications are written on these bags, as this greatly
simplifies further more detailed studies in the future. Bags
were boxed according to site. Particular items such as artefacts
were washed and catalogued in a book, beginning at Ka7 9.1. Next,
each paper bag was tipped into a one-quarter inch(6.35mm) mesh
sieve, and the retained material was sorted with tweezers in
a large plastic tray into:
Bone
Charcoal
Pumice, stone etc
Shell
Artefacts and worked material
and then it was rebagged, again in paper bags. The residue was
also rebagged and kept. Bone material was then sorted into the
main groups(fish, rat, bird, etc), and rebagged. The methods used
in processing individual components are described in the
various Appendices attached.
Fish
The fish remains form a large proportion of the midden
material recovered from Kapingamarangi. The method of
identification is described in Appendix 3, and need not be
enlarged upon here. A minimum number of 1214 fish were identified,
about equal numbers being in Tiroki(KA1),Putau(KA3), and
Ngeiho-Hereu(KA4), and only about one third of this number in
Muri-Harau(KA2). The results can be grouped according to fish
families, and also according to time periods, following the
chronology described above. These reorganised results are given
in Table 5.
About 30 species of fish were identified, and another
three to five species are present in the midden in small numbers
but are not in the comparative collection. As might be expected,
only a few species are ever very numerous. The information is
pooled and reorganised further in Table 6. From this, it will
be seen that fishes in the families Epinephelidae (cod and
groper family), and Scaridae(parrotfish) are equally common.
These are followed by Order Anguilliformes(almost entirely
family Muraenidae, the moray eels), and families Balistidae
(triggerfish), Carangidae(trevally etc), Nemipteridae(entirely
Monotaxis
gvanoculis),
Lutjanidae(snapper etc), and Holocentridae
(squirrel fishes).
TABLE 5. MINIMUM NUMBERS OF FISH, ORGANISED INTO FAMILIES AND CK :RONOLOGICAL PERIODS
Site
Period
Class Elasmobranchii
Order Anguilliformes
Belonidae
Exocoetidae
Holocentridae
Sphyraenidae
Acanthocybiidae
Thunnidae/Katsuwonidae
Carangidae
Epinephelidae
Lutjanidae
Caesiodidae
Nemipteridae
Lethrinidae
Mullidae
Kyphosidae
Coridae
Scaridae
Siganidae
Acanthur idae
Scorpaenidae
Diodontidae
Tetrodontidae
Balistidae
Total
I II III
1
9
11
KA1
IV
I
%
2
10
12
21
3.06
5.36
9 17
1
1
1
1 2
KA2
III IV
1
2
5
9
£
%
6 5.83
11 10.68
KA3
II III IV
3
1
9
18
5 1
13 7
1
3
1
11
1
6.63
0.51
1
5
8
57
5
14 3.57
25 6.38
97 24.74
15 3.83
1 2
1 8
2 21
4
3 2.91
9 8.74
23 22.33
4 3.88
2
2
3
1
16
18
40
10
18
4
1
25
10
2
6.38
2.55
0.51
2 10
1
12 11.65
1 0.97
1
1
13
11
2
1
1
9
17
40
9
7
6
1
3
33
1
4
3
54
26
2
2
2 12
9
6 1.53
87 22.19
1 0.26
13 3.32
1
2
2
7
4
10
1.02
2.55
1
8
11
22
5.61
1
11 22 162 197 392
1
I
0.97
2 1.94
14 13.59
2
2
3.32
4
5
4.85
9
10
13 90 103
9.71
1
9
5
55
1
3
1
5
1
28
2
28 250
2
3
4
4
10
8
1
3
1
1
5
4
4
2
7
2
6
1
l
%
18 4.63
39 10.03
5
1
16
1.29
0.26
4.11
23 5.91
29 .7.46
57 14.65
23 5.91
1 0.26
17 4.37
14 3.60
2 0.51
3
5
11 2.83
76 19.54
2
3 0.77
1 0.26
10 2.57
1 0.26
42 10.80
6
68 43 389
KA4
II III IV
10
30
4
12
2
5
2
6
13
9
35
18
14
5
1
8
26
3
6
11
197
£
%
16 4.85
47 14.24
2
0.61
6
19
5.76
3
2
8
4
1 0.30
21 6.36
15 4.55
54 16.36
22 6.67
2
3
1
1
3 3
18 10
1
1
19 5.76
10 3.03
1 0.30
2 0.61
14 4.24
54 16.36
1 0.30
4 1.21
7
1
5
4
11
3
2
1
4
2
4
7
2
19
76 57 330
2.12
0.61
5.76
Total
52
118
7
1
62
2
1
61
78
231
64
1
73
35
5
2
33
231
2
22
1
26
13
93
1214
I
00
I
-59TABLE 6: FISH REMAINS ARRANGED IN ORDER OF DECREASING ABUNDANCE
I
Ill
no. %
no.
75 15.99
63 19.75
90 23.26
231
9 23.08
81 17.27
60 18.81
81 20.93
231
Anguilliformes
1
2.56
48 10.23
38 11.91
31
8.01
118
Balistidae
3
7.69
41
8.74
19
5.96
30
7.75
93
Carangidae
2
5.13
27
5.76
26
8.15
23
5.94
78
Nemipteridae
1
2.56
27
5.76
15
4.70
30
7.75
73
Lutjanidae
1
2.56
29
6.18
17
5.33
17
4.39
64
Holocentridae
9 23.08
34
7.25
9
2.82
10
2.58
62
Vevuod
II
no.
%
Epinephelidae
3
7.69
Scaridae
no.
%
IV
%
Total
Thunnidae/Katsuwonidae 2
5.13
29
6.18
19
5.96
11
0.26
61
Elasmobranchii
3
7.69
20
4.26
19
5.96
10
2.58
52
Lethrinidae
1
2.56
16
3.41
9
2.82
9
2.33
35
Coridae
1
2.56
13
2.77
8
2.51
11
2.84
33
Diodontidae
2
5.13
11
2.35
5
1.57
8
2.07
26
Acanthuridae
6
1.28
4
1.25
12
3.10
22
Tetrodontidae
2
0.43
2
0.63
9
2.33
13
5
1.07
1
0.26
7
2
0.43
Belonidae
1
2.56
Mullidae
Sphyraenidae
Kyphosidae
1
1
0.31
2
0.52
5
1
0.31
1
0.26
2
1
0.26
2
0.21
Siganidae
2
0.63
2
Scorpaenidae
1
0.21
1
Exocoetidae
1
0.21
1
Acanthocybiidae
1
0.31
1
Caesiodidae
1
0.31
1
Total
39
Shannon's H
(bits)
Volume
Fish
statistic
(m3)
density
MN/m
3
469
319
387
1214
3.30
3. 65
3. 53
3. 46
5. 20
20. 20
19. 20
9.60
54.20
7. 50
23. 22-
16. 61
40. 31
22.40
-60Fish in these eight most common groups account for over 78% of
the total. Of particular significance in addition to these eight
are specimens of Thunnidae and Katsuwonidae, accounting for
just over 5% of the entire catch(Minimum Number,MN=61). These
are the main representatives of the pelagic game fish, primarily
caught with lure hooks, trolled from canoes. Other fish which
would probably have been caught in this way are
Aoanthooybium
(MN=1), family Sphyraenidae(MN=2), and
Elegatis
solandri
(MN=5). Taken together with the tuna families,
bipinnulatus
this type of fish represents 5.68% of the total catch - a small
proportion compared with the great importance attached to these
fish today on Kapingamarangi. This feature is returned to later.
Three overall observations should be made at this stage.
Firstly, it will be seen from Table 6 that the general
abundance of fish in the archaeological sites increases more
than five-fold during the sequence(7.5, 23.2, 16.6, 40.3 MN/m 3 ).
Even taking into account the fact that the sediments forming
Touhou accumulated at different rates, this dramatic increase
must partly reflect increasing population size on the islet.
Secondly, there are some significant changes through time
in the relative abundance of different types of fish. For
instance,the clearest trend is in the cod and groper family
( Epinephelidae) , which increases some threefold in its relative
importance(7.7,16.0,19.8,23.3% of the catches at each period).
By contrast, the parrotfish(Scaridae) are by far the most
numerous early in the sequence(23.1%), and their importance
falls somewhat through time, though not uniformly. Moray eels
are not important early in the sequence, but increase in
popularity.
Thirdly, there could be an' areal factor relating to site
function, which complicates the task of interpreting changes
through time. From Table 5 it will be seen that the main fish
types are not uniformly abundant from one site to another.
A number of different reasons could be advanced for the observed
disuniformity. Different families would not necessarily have
equal access to the different fishing areas in the atoll,
nor would all families have fishermen with equal skills in all
kinds of fishing methods. One therefore has to be careful not
to interpret too closely the trends observed, because this
factor is difficult to disentangle. Pooling results helps to
minimise this effect.
As for catching methods, the main fish types would have
been caught by rather different means. The cod and groper
family, although caught in the lagoon, are normally in somewhat
deeper water, and would have been primarily caught with baited
'hook. The parrotfish, on the other hand, can be taken in quite
shallow water around coral thickets, but since their diet is
chiefly vegetable matter, invertebrates, molluscs, and coral,
they are only rarely caught on a baited hook. These fish can
be speared of course, but above all they must have been taken
with various kinds of nets. The eels represented could have
been caught by a number of different methods too, but given
the elaborate trapping methods used by Kapingamarangi
-61fishermen in the historic era for catching eels, this
interpretation is preferred. In the time trends observed above,
therefore, we should also read into the patterns some changes
in fishing technology. In particular, netting seems to have
dominated the early part of the sequence, and baited hook
techniques appear to increase progressively in importance.
Trap fishing seems initially fairly rare, but also gains
in popularity.
The pattern of relative abundance of the different kinds
of fish was assessed using Shannon's H statistic(Table 6 ) , and
the results are fairly uniform for the different periods.
By comparison with the statistics for shellfish(see below),
the values obtained indicate a more even emphasis on a greater
number of target species.
Shellfish
The shellfish recovered in the excavations were identified
to species level wherever possible, and minimum numbers
calculated. The details of this process are given in Appendix 5.
More than 70 species were recovered, but of these only a few
can be considered to have had any real economic significance.
A summary of the shellfish remains is given in Table 7. It will
be readily apparent from this that all but about 13% can be
accounted for by shells belonging to three families. By far
the most common shellfish is Turbo argyrostomus
(nearly 68%),
followed by Donax sp.(14.5%), and Cypraea spp.(6.9%). Only 59
specimens of Tridaona
cf maxima were found, and only eight of
Hippopus
hippopus.
It is very likely that the bulk of these
latter two shells were on site for industrial reasons, rather
than as food. Today, the esteemed food of Tridaona
maxima is
gathered on the reef, and the shells discarded on the spot
as the meat is separated. Although the modern inhabitants have
the benefit of steel knives for this purpose, this pattern of
exploitation must have prevailed in antiquity too. It is quite
likely that during the prehistoric period, this species
constituted the most important component of the shellfish
diet - this is not reflected in this Table of results. Hippopus
hippopus,
on the other hand, is somewhat more difficult to
deflesh without a fire, and is more likely to have been brought
home if it was eaten at all. Their small number in the sites
probably does reflect their unimportance in the diet, and they
are regarded with distaste today. As pointed out in Appendix 5,
most of the specimens of Cypraea spp. were very small, and few
can be considered to have been items of food. Specimens of
Donax sp. too, could easily have been imported
in the
kilikili
gravel; again, their economic significance is questionable.
Finally, shells of Turbo sp. are a favourite home for hermit
crabs, and some specimens must have been imported by them
from the surrounding beaches. Even so, concentrations of these
shells in the sites strongly suggest that the bulk of these
were indeed collected by people for food. It is interesting that
neither Nerita
spp. nor Donax sp. figure prominently in the diet
TABLE 7: MINIMUM NUMBERS OF SHELLFISH ARRANGED IN FAMILIES AND CHRONOLOGICAL PERIODS
Period
Site
Trochidae
Turbinidae
Neritidae
Cerithiidae
Strombidae
Cypraeidae
Naticidae
Cassidae
Cymatiidae
Bursidae
Muricidae
Magilidae
Buccinidae
Vasidae
Harpidae
Mitridae
Conidae
Terebridae
Ellobiidae
Arcidae
Pinnidae
Pteriidae
Pectinidae
Spondylidae.
Carditidae
Cardiidae
Tridacnidae
Mesodesmatidae
Donacidae
Tellinidae
Psammobiidae
Trapeziidae
Veneridae
IV
KA1 KA2 KA3 KA4
2
22 401
3 49
2
4 21
4 262
1
1
1
1
1
E
2
56
5
1
1
68 547
57
2
26
267
1
1
1
4
4
%
III
KA1 KA2 KA3 KA4
0.14
37.19 301
3.87
6
0.14
1
1.77 15
18.15
4
0.07
0.07
0.07
5
5
0.34
0.34
6
17
5
0.41
1.16
0.34
1
1
1
2
1
2
1
1
1
6
1
1
5
7
3
1
6
4
5
4
7
1.43
0.07
0.14
3
2
(bits)
4
12
1
11
18
0.75
1.22
458
1
2
1
8
31.14
0.07
0.14
0.07
0.54
1471
2. 36
17
5
1
0.41
6
8 450
1
2
1
8
Totals
Shannon's H statistic
21
1
2
%
II
KAl KA3 KA4
I
E
%
KAl KA3
E
%
1
1 0.07
1
11 0.37 1
4 0.64
10
3
52 139 559 1051 73.81 181 860 1147 2188 73.65 12 472 484 77.07
5
21 1.47
4 40 32
8
2
76 2.56 2 13 15 2.39
1 0.07
3
1
1 0.16
3
6 0.20
98 3.30
1 0.16
16 100
3 134 9.41
19 79
1
13 12
38 2.67 10 38 59 107 3.57 2 32 34 5.41
9
4
1
5 0.17
1
1 0.16
1
6
17
5
E
3
1
2
4
1
4
5
2
25
56
7
1
1
1
1
1
2
1
0.07
0.07
0.07
0.14
0.07
1
2
1
1
1
34
1
1
0.07
0.14
0.07
0.07
0.07
2.39
0.07
0.07
1
2
1
2
2
1
1
1
1
3
3
14
3
10
1
14
5
2
1
3
1
7
6
3
3
1
14
13
0.21
0.07
0.98
0.91
9
97
6.81
1
0.07
6
1 13
1
1
34 199
1
1 12
1
0.07
5
1424
1. 55
2 .
4
1
2
14
95
1
5
4
5 0.17
1 0.03
2 0.07
3 0.10
1
1
4
27
3
0.03
0.03
0.13
0.91
0.10
20
0.67
5 0.17
1 0.03
12 0.40
1 0.03
8 0.27
28 0.94
2 0.07
328 11.04
2 0.07
18 0.61
9
0.30
2972
1. 61
2
2
1
2
2
0.32
1
6
1
6
0.16
0.96
7
7
1.11
1
6
1
8
0.16
1.27
55
1
3
57
1
4
9.08
0.16
0.64
1
1
0.16
628
1. 40
Total
%
18 0.28
4270 65.75
169 2.60
10 0.15
259 3.99
446 6.85
7 0.11
1 0.02
9 0.14
2 0.03
8 0.12
10 0.15
1 0.02
1 0.02
2 0.03 i
13 0.20 £
51 0.79 l
9 0.14
1 0.02
75 1.15
2 0.03
8 0.12
1 0.02
28 0.43
2 0.03
34 0.52
67 1.03
2 0.03
940 14.47
4 0.06
25 0.38
1 0.02
19 0.29
6495
-63at any time during the prehistoric period. This is a striking
contrast to archaeological remains in some sites in the
Solomon Islands, where these shells appear to increase in
importance, as larger species become locally depleted through
exploitation(see Leach and Davidson,n.d.).
Despite the fact that only a few species are economically
important, or perhaps because of it, there are some interesting
patterns which can be observed through time. The first feature
which can be noted is that the general pattern of shellfish
exploitation is quite different in the last 100 years(Period IV),
than anything in the previous centuries. This is most clearly
seen by the values of Shannon's H statistic for the four periods
(see Table 7 ) . In ecological studies, this is used to measure
species diversity against environmental stress(Wilhm,1968).
The value of H=2.36 for Period IV shows that the pattern of
diversity is markedly less than in the prehistoric period, where
an average of about H=l.5 prevails. This is largely due to near
equal quantities of Turbo sp. and Donax sp.(37.2 and 31.1%
respectively), and high proportions of Cypraea spp. also(18.2%).
In earlier periods, by contrast, Turbo sp. occupies a position
of far greater importance, and the scree of percentages from
the level for this species falls off much more steeply. In
economic terms this implies that although the prehistoric
people collected a wide range of species, their prime targets
were few in number, and illustrates a highly conservative
approach to shellfish gathering. As pointed out above, the
pattern of exploitation of fish, on the other hand, is quite
different, and shows greater enterprise.
Leaving aside this apparently major shift in emphasis in
the last 100 years, some more subtle trends can be observed in
the earlier levels. The figures for Turbo sp. decrease somewhat
(77.1 to 73.6 to 73.8%), as do those for Cypraea spp.(5.4 to 3.6
to 2.7%). The main contrasting shellfish making up the difference
are specimens of the family Strombidae which rise in frequency
(0.2 to 3.3 to 9.4%). This might suggest that two of the
main target types of shellfish were becoming harder to obtain
in the course of time.
A total of 6494 shellfish in 54.2 m 3 of excavated deposits
is not a high density in comparison with some occupation layers
in other Pacific islands, and with the possible exception of
Tridacna
maxima,
shellfish do not seem to have been of major
significance in the early diet on Kapingamarangi.
Crustaceans
The analysis of crustaceans is given in Appendix 4. Only
a few fragments were recovered in the excavations, although
conditions are considered favourable for their preservation
(see Leach and Anderson,1979 b) .This is taken to be good evidence
that crustaceans were not very important as an item of
prehistoric diet.
'-64Crayfish are represented by only two mandibles in
Level II, and coconut crab is only found in Level IV.
Unfortunately, very little is known of the spread of the
coconut crab through the Pacific, but this result should not
necessarily be taken to indicate their absence on Kapingamarangi
prior to the last 100 years. The exoskeleton of this animal
appears to be richer in protein than many other crustaceans,
and their archaeological survival is therefore much less
certain. The lack of crayfish though is rather puzzling; however,
neither of the two species found on Kapingamarangi today figures
prominently in the modern diet either.
Birds.
About 70 fragments of bird bones were recovered in the
excavations - a very small number compared with some Pacific
island sites. Sufficient of these could be identified to show
that the frigate bird, red-footed booby, and noddy tern were
definitely items of prehistoric diet(Table 8 ) . These all appear
early in the sequence in Level II. The chicken is certainly
present by Level IV. Bones of Gallus
sp. were recovered in close
association with the human skeletal material described in.
Appendix 10. The age of this material is unknown, and the
chicken bones were unfortunately lost in transit. The human
bones are thought to be definitely prehistoric, and it is
possible the chicken bones were too - the tarsometatarsus had
a strongly developed spur, characteristic of the prehistoric
chicken, and not so frequently present with modern chickens.
sp. have been identified from Nukuoro.
No bones of Gallus
TABLE 8: BIRD MINIMUM NUMBERS FROM THE EXCAVATIONS
Period
Site
Bird
iv
III
II
i
KA1 KA2 KA3 KA4 KA1 KA2 KA3 KA4 KAl KA3 KA4 KAl KA3
? sp.
1
Gallus gallus
Fregata
minor
1
1
7
3
3
subsp.
6
1
An&us stolidus piliatus
Sula sula rubripes
1
2
1
-65Pig, Rat, Turtle, and Human
The minimum numbers of these are listed in Table 9. With
the exception of rat, the bones of these animals were
relatively rare. It is notable that not one fragment of dog
was found. During the excavation at Putau(KA3) a pig's tooth
was tentatively identified deep in Level II in a prehistoric
context. This important find unfortunately was lost in the
freight, and cannot now be confirmed. It is possible that the
tooth was actually from a specimen of
Pseudobalistes
flavimavginatus.
This fish grows to a large size on Kapingamarangi,
and some of the teeth have a superficial resemblance to pig
incisors.
All remains of rat were certainly of Rattus
exulans,
and
are quite frequent relatively early in the sequence. It is
notable that none were found in Level I. This reinforces the
view that Touhou islet at this time was barely above the sea,
unsuitable for permanent human habitation, and had not been
colonised by rat at this time.
The general rarity of turtle bones is interesting, and
probably shows that the sea grass flats of Zostera
sp. around
Kapingamarangi were never more extensive than they are today.
Only one human tooth was found in the excavations. It is
unusual not to find small pieces of human bone and teeth in
Polynesian archaeological sites.
TABLE 9: MINIMUM NUMBERS OF TURTLE AND VARIOUS MAMMALS IN THE EXCAVATIONS
Period
Site
Sus
sarofa
IV
KAl KA2
KA KA3 KA4
KA1
1
Turtle
2
Homo sapiens
Rattus exulans
Mammal
3
? sp.
* = 1 tooth
III
KAl KA2 KA3 KA4
II
KAl KA3 KA4
I
KAl KA3
Total
1
1
5
1
1
4
1
10
1*
2
3
2
3
1
7
1
4
1 6
17
20
1
64
1
4
-66Stone and Pumice
Some stone material was recovered, but it is not thought
that all was retrieved by any means. Without washing on the
spot, it can be very difficult to distinguish between dense
burnt coral and stone. When the lowest levels of Putau(KA3)
were excavated, it was well below sea level, and buckets
containing a slurry were wet sieved in the sea. This revealed
some most interesting foreign rocks(see Appendix 9 ) . In addition,
pumice was seen at various levels in all sites, and although
samples were taken, this was not systematic. On the whole,
pumice seemed to occur as concentrations at different levels,
and may well reflect the periodic arrival of significant
quantities on beaches, corresponding with activity from a distant
volcano. Of course, this pumice is not now in primary deposition,
since it was imported onto Touhou with the gravel material
while building up the islet. Its historical significance would
therefore be very difficult to establish.
The rocks in the sites were almost certainly derived
from the roots of drifting trees, which still regularly arrive
on the island. In fact, several of the stones described in
Appendix 9 were taken from driftwood during the expedition.
Ultimately it may prove possible to identify the sources of
these stones, and this may be useful information to assess
the possible areas from which drift voyages took place to
Kapingamarangi during the prehistoric period. The occurrence
of such drifted stone should also serve to introduce a note
of caution in interpreting possible exchange networks of
such things as oven stones in Pacific archaeology.
HUMAN SKELETAL MATERIAL
Human bone, representing six individuals,was collected
by Lieber from the foreshore after it eroded from the section
immediately to the north of the excavation at Putau(KA3) following
a storm and exceptional tides. Unfortunately the remains
are not directly associated with archaeolgical sediments, and
their age is therefore uncertain. However, they are certainly
of some antiquity, as burial plots in this vicinity are not
known by the modern inhabitants. It is fairly safe to assume
therefore that they are prehistoric in age. The material has
been studied by Houghton(see Appendix 10), who clearly
identifies the people as Polynesian in their biological
relationships. There is no sign of yaws, and the people appear
to have been well nourished.
-67ARTEFACTS
Shell Adzes
A total of 23 shell adzes were found. Only 12 of these
were recovered in controlled excavations, the remainder were
surface finds(see Figures 37 and 38). The shell species cannot
be accurately determined for any, although Tridaana
cf.maxima
is probable for all. Some of the adzes are quite small in size,
although there is no chronological significance which can be
attached to this. Specimens of both large and small(Figure 37:
#B, #J) were found in Level I at Putau(KA3), showing the presence
of both forms early in the sequence. One adze is unusual(Figure
37:#H) in having a definite waisted appearance, otherwise these
adzes are perfectly ordinary in the general context of island
cultures anywhere in the Pacific. It is possible that the waisted
specimen is a coconut grating head(see below).
During the Bishop Museum expedition, 2 9 shell adzes
were recovered(Buck,1950:165); these were all of Tridaana
shell,
and most were the common ventral lip form. A few are of the
thicker variety made from the solid interior, as is probably
a specimen figured here(Figure 37,#L). One other borderline
example was also found, but is not figured. It is interesting
that Buck also collected a Cassis
sp. chisel, but could not
obtain information as to its use(Buck,1950:172). This type of
tool is absent on Nukuoro(Davidson,1971:68).
The single most important feature of the adze assemblage
is the total lack of adzes made from Mitra sp. or Terebra
sp.
It is possible that this adze form was manufactured in the past,
and that the excavations simply failed to produce any examples.
One point in favour of this view is that an old informant when
questioned about the possibility of such adzes, proffered the
name toki poko for those made from Mitra sp.( pu manga roro) .
Lieber has identified this term as that applied to gouging
adzes(Lieber and Dikepa, 1974), and this strengthens the case.
However, it is also clear from discussions with old men that
considerable influence from Nukuoro has been felt in adze
technology, particularly in methods of hafting, during the
historic era; it is therefore possible that this is where the
knowledge of this adze form derives. On Nukuoro, adzes of either
Terebra
sp. or of Mitra sp. are quite numerous(Davidson,1971:52).
A comparison between the two assemblages is given in Table 10.
One is forced to the conclusion that if adzes of
Mitra/Terebra
were manufactured in the past on Kapingamarangi, they were
extremely rare. Suitable large specimens of these shells are
quite common on the island. This matter is of some potential
importance, since one of the few aspects, of culture-historical
significance which has been found amongst the different types
of shell adze in the Pacific, is that this form of adze
developed relatively late in time. In Taumako, in the Solomon
islands, for example, an antiquity of no more than 1000 years
is suspected. The relative abundance of this adze type on
-68-
i
i
i — i
i
i
cm
F i g u r e 37: Tvtdaona s h e l l adzes from Kapingamarangi. The provenances a r e as
f o l l o w s : KAl:A(Spit 3 ) , G ( S p i t 8);KA2:D & E(Layer 5-6);KA3:B(Spit 2 1 ) , C ( S p i t 9 ) ,
J ( S p i t 20),M(Spit 7 ) ; KA4:I(Spit l l ) , K ( S p i t 1 2 ) , N ( S p i t 8 ) , 0 ( S p i t 6 ) ;
Surface f i n d s : F & H ( P u t a u ) , L(Werua).
-69-
Figure 38: Surface collected Tridaana
shell adzes from Kapingamarangi.
TABLE 10: COMPARISON OF ADZE FORMS FROM NUKUORO AND KAPINGAMARANGI
(see Davidson, 1971: 52)
Nukuoro
Terebra
Mitra
Cassis
maoulata
sp.
Kapingamarangi
69
2
sp.
5
1
Tridacna maxima
(ventral lip)
68
23
Tridacna maxima
(solid interior)
21
Total
165
24
Note: The Cassis sp. example from Kapingamarangi is a coconut grater, as
perhaps the Nukuoro specimens are too.
Buck (1950: 17 2) records the Cassis sp. chisel on Kapingamarangi,
but it is not known on Nukuoro.
-70Nukuoro, and their presence in the earliest levels here suggests
that the island was settled from some area where the form had
already been developed. In Kapingamarangi, on the other hand,
their apparent absence through the sequence might indicate
an origin for the founding group from some area where the
adze was not known, or not yet developed. The striking difference
in character between the adze assemblages from Kapingamarangi
and its nearest neighbour Nukuoro is compelling evidence of
cultural isolation through the prehistoric period, and perhaps
indicates different origins for the two founding groups.
One-Piece Fish Hooks
The archaeological sites were practically bereft of evidence
of fish hook manufacture, considering the quantity of material
excavated. The few examples recovered are illustrated in Figure
39, and the debitage from manufacture in Figure 44. There are
some interesting features of this small collection. Of the
three definitely prehistoric hooks found, two are made of
bone(possibly turtle), and the other is pearl shell. The three
are bracketed by two radiocarbon dates(NZ5145 and NZ5146) to
between 350 and 550 years BP(See Figure 33). Most of the few
fragments of worked pearl shell were found in the same provenance
(Figure 44). The authors saw more fragments of worked pearl
shell scattered over the surface of Nukuoro in a few hours,
than in the sustained archaeological work on Kapingamarangi.
This dearth of shell fish hook material in the sites
unquestionably reflects the extreme rarity of Pinotada
spp.,
and of Atvina
vexillum
in the Kapingamarangi lagoon. Clearly,
some hooks were made from pearl shell when it was available,
but if fish hooks were common in the past they must have been
made from some less durable material which has not survived
archaeologically. The large number of fish of the family
Epinephelidae in the sites could really only have been taken
with baited hooks - most must have been made from wood, turtle
shell, bone or coconut shell. The information given by Eilers
on fish hooks(1934:73ff) is instructive in this respect(Buck's
comments[1950:231] are largely a repetition from Eilers).
Small hooks are said to have been made from coconut shell,
turtle shell, pearl shell, and black mussel. Large hooks were
said to have been made from wood, pearl shell, and black mussel.
Of these, in our opinion, coconut shell must have been
especially common; and this view is supported by the illustration
in Eilers1 work of 16 such hooks on a special hook holder
(1934:74; see also Buck,1950:240). This view is also held by
Anell(1955:95), although apart from a few more recently collected
items, again he largely relies on Eilers' earlier observations.
In passing, an error should be noted in Anell. Specimen #6 in his
Figure 4 from Kapingamarangi, is said to be of pearl shell,
and specimen #7 of coconut shell, following Eilers (1934:94) .
In fact, the specimens are cited(Eilers,1934:75,76) as being
made of pearl and black mussel respectively. Beasley(1928:101)
is able to add very little to the subject, as his observations
for Kapingamarangi depend on a single and rather crude wooden
hook collected by Commander Erskine about 1850. Turtle bone
fish hooks are relatively rare in the Pacific islands(some are
t
•
i
i
i
i
ic m
Figure 39: One-piece fish hooks from Kapingamarangi. They are made from bone
(B,E), metal(F,G), Pinatada sp.(A,c,D,I), and Atrina vexillum(U).
Both metal
hooks are from KA2(Layer 4), the two bone hooks are from KA4(B from Spit 10,
and E from Spit 13). Specimen A is from KA4(Spit 10),D was found while digging
a grave on Werua,C and I were surface finds from the eroding section at
Putau(KA3), and H is a modern handicraft item, made in 1980.
•-72known from Funafuti[Hedley,1897:266]), and again their presence
on Kapingamarangi attests the difficulty of finding suitable
material for hook manufacture, as turtle bone is not very strong.
Three types of hook are evident in the small collection.
Firstly, the plain rotating hook with slightly incurved point
is represented by specimens A,B,C, and E(Figure 39). The method
of line attachement obviously varied, as one specimen is notched
externally on the upper shank, another is plain, and yet
another has a slight inner projection. Secondly, specimen D has
a short pointed shank, and a markedly incurved point. The
example is notched at the base of the shank and has several
grooves along the lower margin. These features are closely in
line with the method of snooding illustrated by Eilers (1934:75,
specimen #22). This hook was found by an old man while digging
a grave on Werua, and he was unequivocal in his name for the
hook - matau kina - the hook for the rainbow runner,
Elegatis
bipinnulatus.
A hook type with an equally incurved point from
Nukuoro(Type VII) is claimed to have been used for catching
an important fish called gina (Davidson,1967:187), which may
well be the same species.Thirdly, specimen I is unusual in
possessing a snooding knob on the inner proximal end of the
shank. This is almost a perfect match of the hook illustrated
by Eilers(1934:76,#27), and copied by Anell(1955:93,#6). It
has a fairly fresh appearance, and is not thought to be very
ancient. It is a rather crude shape, with only bare attention
given to final finishing. In our opinion it is unlikely to
have been useful for actually catching fish. It is quite likely
to have been a handicraft item to satisfy the voracious collecting
habits of Europeans at any time in the last century. Eilers
comments that souvenir manufacture was firmly established on
Nukuoro by 1910(1934:229). It would be wise to wait until this
particular hook form was found in a firm prehistoric context
before commenting on possible parallels with inner shank knobbed
hooks from elsewhere in the Caroline islands(see Anell,1955:92).
The Kapingamarangi people are very quick to learn new shapes
from books, which they then turn out as handicraft items for
sale to modern collectors. It should be noted that Buck showed
the fish hook illustrations in Eilers to people in 1947(Buck,
1950:232), and they showed him how to reproduce the lashing from
these illustrations. It would not be at all surprising if specimen
I illustrated here was not a direct copy of the one illustrated
by Eilers. Where the original idea may have come from is
open to conjecture. The final stages of this process of
morphological degeneration in the face of continuing European
demand for curios stripped of their functional significance is
evident from specimen H. This was made with steel tools by an
old man in 1980 from Atrina
vexillum,
and has fibreglass cordage.
Finally, two metal hooks were recovered in the Muri-Harau
excavation. One of these is shaped in typical Polynesian style,
and is made of strong brass or bronze; the other is made of softer
coated brass, and is quite a crude shape.
-73Trolling Lure Shanks
Shanks for trolling lures are stylistically an important
artefact from Kapingamarangi. Four specimens are illustrated
in Figure 40. Three are of dense white shell, possibly
Tvidacna
sp., but it could be Spondylus
vavians (specimens A,B,C). The
fourth is of Pinatada
sp.(specimen D ) . These shanks are very
finely made, and are distinctive in having a groove running
nearly the full length of the dorsal surface of the shank,
although in the case of the pearl shell example this groove is
much shorter. Portions of two other pearl shell lure shanks were
also found. One came from the gravel lens towards the base of
Layer 4 in the Putau site(KA3, see Figure 30), and the other was
found in the eroding section north of Putau(KA3). These were
i
i
i
1
i
1
i crn
Figure 40:Shell lure shanks from Kapingamarangi.A:found while digging a grave
in the Ngeiho-Hereu area(KA4). B,C:eroded from scarp at Putau area(KA3).
D:KAl,Spit 15, next to radiocarbon sample NZ5135.A11 have the characteristic
groove along the dorsal surface, unique to Kapingamarangi.
-74unfortunately lost in the freight transport. Both also had
relatively short grooves, and were comparable to specimen D.
Specimens B and C are a rather rounded triangular shape in cross
section, but #A is distinctly sharp-edged with a sub-triangular
section. Buck saw one similar shank during his visit to the
island(1950:237), and another is in the Bishop Museum(C.10170).
In addition, Buck regarded a set of 9 very similar items which
he recovered as a necklace(1950:273ff). These do not have the
characteristic dorsal groove, although they appear identical
in all other respects. Eilers illustrates a pearl shell lure
(1934:77), and vividly describes the dorsal grooving, likening
the distal end to a fishtail. The barbed point is made of turtle
shell. The crude untrimmed nature of the proximal end of this
specimen led Buck to comment on its Marshallese affinities(1950:
238); something which some authors believe was quite deliberate
on the part of the artisans, as it is thought to enhance the
fish luring qualities(see Beasley,1928:95;Anell,1955:152). It is
possible that this feature was introduced to Kapingamarangi
during the Marshallese invasion of about 1870. As for the
grooved feature on these Kapingamarangi lures, no cultural
parallel can be suggested. The pearl shell versions, although
not fully developed in style, establish the form as thoroughly
diagnostic of Kapingamarangi, with a respectable antiquity of
about 700 years. The lure shanks recovered archaeologically
on Nukuoro, are of more widely recognised types.
Coconut Grater Heads
Three specimens of these are illustrated in Figure 41. The
one made from Cassis
sp is bracketed by two radiocarbon samples
(NZ5141 and NZ5140), and has an age of about 600 years. The
second example(B) is also of shell, but the species is obscure.
It was possibly made from a juvenile specimen of Spondylus
vavians.
This one is closely associated with another radiocarbon sample
(NZ5145), and has an age of about 350 years. Finally, the third
less certain example(C) is made from Atrina
cf, vexillum
, and
is a surface find. It is interesting that no clear evidence for
bivalve coconut hand scrapers was seen. This is surprising in
view of their very common occurrence in islands to the south,
V
A-A_>}|_£_A_/i. _* J
Figure 41: Shell coconut grater heads from Kapingamarangi. A-.Cassis sp. from
KA3(Spit 14). B: from KA4(Spit 9). C: Atrina cf,vexillum surface find near
Putau(KA3).
-75and presence on Nukuoro. Non-bivalve coconut grater heads on
Nukuoro, although common, appear to come into the archaeological
record relatively late, possibly no earlier than 300 years
(Davidson,1971:72); they are at least twice this age on
Kapingamarangi.
Buck collected a number of adze-like objects made of
Tridacna
sp. which old men recognised as coconut grater heads
for stools. The main distinguishing feature is the constricted
sides for lashing onto the stool. It is possible that one of
the adzes illustrated here(Figure 37,#H) really functioned
in this manner. In addition, Eilers illustrates a stool(1934:122)
with a bivalve shell attached, identified as Cardium sp.
Pottery
A single sherd of what is thought to be prehistoric
Oceanic pottery was found on the surface of the Tiroki site(KAl)
immediately before excavation. This is the subject of three
technical reports(Appendices 6,7, and 8) which examine its
physical and mineralogical properties. The sherd is illustrated
in Figure 42(#E). It is difficult to arrive at a hard and fast
conclusion about this sherd. One possibility which was considered
was that it was a fragment of a Japanese field tile(dokan).
These are actually quite common in Micronesia, but should be
readily distinguished from their curvature and glazing on both
surfaces. To be thoroughly certain, however, samples of these
tiles from Kapingamarangi were subjected to mineralogical
investigation(Appendix 7 ) . The sherd is quite different in
composition to these tiles. The mineralogy of Spanish pottery
from Pacific contexts(see Leach and Davidson,n.d.) is again
quite different. In general physical appearance it is quite
similar to prehistoric pottery to the south and east, but it is
exceptionally hard. Since it was found on an atoll, we can
be quite certain it was imported; and the fact that it was not
found deeply buried in an archaeological site does not rule
out the possibility that it arrived with very early visitors
to the island, perhaps with the founding group. Of course it
may also have been imported during the historic era. Indirect
contact with populations to the south was reasonably frequent
during the period of German administration(1899-1914), and
the sherd does resemble the well-fired modern trade pottery
from the Siassi area of New Britain, for example. On the other
hand, the mineralogy is consistent with(but not proof of) a
derivation from the Solomons area.
Both in physical properties, and mineralogy, the sherd
is quite distinctive from known Micronesian potteries.
I
I
I
I
'
I
cms
Figure 42: Ceramic and metal a r t e f a c t s from Kapingamarangi. Item E i s a
s u r f a c e c o l l e c t e d pot sherd from t h e T i r o k i s i t e ( K A l ) . The remainder a r e
from t h e Japanese s e t t l e m e n t a r e a s on Nunakita and P a r a k a h i . Aicopper o b j e c t .
B:Tin bronze i n c e n s e burner.C:50mm US Browning a i r c r a f t machine gun c a s e .
D: Lead o b j e c t . E: Pot s h e r d . F , G , H , I : Japanese g i a z e d f i e l d t i l e s ( d o k a n ) .
Shell Arm Rings
Fragments of three b r a c e l e t s or arm rings were found(see
Figure 43). One of these i s made from a large specimen of Conus sp
and i s similar to an example i l l u s t r a t e d by Buck(1950:274), but
much l a r g e r . Buck's specimen has an outside diameter of 4 3mm,
while the archaeological one i s nearer 70mm. Buck regards t h i s
smaller specimen as an ear ornament, but the l a r g e r one would
f i t over a small hand. I t i s about 360 years in age.
The other two rings are most i n t e r e s t i n g . They were
both found in the basal c u l t u r a l layer of Putau(KA3) below
sea l e v e l , and in the t i d a l sea water(see Figure 30). This
context i s believed to be amongst the e a r l i e s t c u l t u r a l
material on t h i s i s l e t , g r e a t e r than 700 years BP. Both
specimens are t r i a n g u l a r in c r o s s - s e c t i o n , and are d e l i c a t e l y
fashioned from a white s h e l l . This i s probably Tridaona s p . ,
but could also be Spondylus vavians . One similar type of ring
was found on Nukuoro in an early context(NU5,Level 16). This
was also t r i a n g u l a r in cross s e c t i o n , and probably of Tridaona sp.
The remaining Nukuoro examples are mainly of a b i v a l v e , such
as Codakia s p . , although one i s thought to be Troohus sp.
Thus, the t r i a n g u l a r - s e c t i o n e d Tridaona sp. arm ring continues
to show i t s e l f to be associated with archaic c u l t u r e s in the
Pacific i s l a n d s . The d e t a i l s of i t s Oceanic development and
spread have yet to be demonstrated.
-77-
i
i
i
i
i
i
. i cm
Figure 43: Shell arm rings from Kapingamarangi. A: Tri-daena sp. 8cm inside
diameter from KA3(Spit 21). B: Tridaona sp. 7cm inside diameter from KA3
(Spit 20). C: Conus sp. 6cm inside diameter from KA4(Spit 9).
-78Perforated Shells and Shell Beads
A considerable number of shells of Nerita
sp. were
perforated; however few if any of these are thought to be
cultural. Some have natural breaks, and many are drilled out
neatly, but are probably the boreholes of Muricid spp. One or
two holes may have been made by grinding on a flat surface such
as a coral slab, and these do suggest human manufacture. The
distinction between human and natural performations is difficult
to be certain about. In Taumako in the Solomon Islands, most of
the shell necklace units had been punched out to form the hole.
The reason why they are certain to have been cultural was
because they were found in situ
around human burials.
It is interesting that although shells of Polinices
spp.
were not uncommon in the sites excavated, none were perforated.
These shells are a favourite for necklaces in islands to the
south, and also in the Kiripati islands. Eilers illustrates
a necklace string of perforated Melampus luteus
from
Kapingamarangi(Eilers,1934:120), and drilled shells are an
important aspect of the modern handicraft industry on the
island.
Only two shell beads were found, and both are from
Ngeiho-Hereu(Spit 13) dating to about 500 years BP. They
cf. ducalis
show flashes of orange-red, characteristic of Spondylus
(see Figure 44). Such beads are widespread in the north Pacific,
including early Nukuoro(Davidson,1971:48), and they also occur
in early and late contexts in various parts of Melanesia.
Drilled Sharks Teeth
Two small shark's teeth which had been neatly drilled were
found in the Putau site(KA3, Spits 8 and 10). The age here is
between 300 and 5 00 years BP. Spears with shark's teeth were
described to Hambruch during his visit in 1910, but none were
actually observed(Eilers,1934:120). It is quite possible that
these were the spears with shark's teeth carried by the
Marshallese castaways of 1870, described in vivid detail to
Hambruch(Eilers,1934:130). Commonly, the presence of drilled
shark's teeth units is taken as an indication of weaponry,
but they have a variety of potential uses on small islands in
the Pacific. For instance, individual teeth were attached to
small sticks and used for tattooing and lancing on Funafuti
(Hedley,1897:300), and Vaitupu(Kennedy,1931:Plate 19). Both
Buck(1950:278) and Eilers(1934:118), however, are adamant
that tattooing was never practiced on Kapingamarangi. Despite this,
tattooing is understood by Thilenius(1903:38) to have been
introduced to Ontong Java from Kapingamarangi. Similar units
were recovered in the excavations on Nukuoro, where incidentally
tattooing was part of the culture, and Davidson prefers a
more mundane interpretation than weaponry also (1971:79) .
-79-
\
i
1
i
i
i
i cm
F i g u r e 44: F i l e s , s h e l l b e a d s , and i n d u s t r i a l p e a r l s h e l l from Kapingamarangi.
The f i l e s a r e made from sea u r c h i n spines.D:KA1(Spit 8 ) . A and B:KA3(Spit 1 2 ) .
C:KA3(Spit 8 ) . The s h e l l beads a r e both of Spondylus cf. ducalis,
and both
a r e from KA4(Spit 1 3 ) . The remaining items a r e p i e c e s of i n d u s t r i a l p e a r l
s h e l l . J and K : s u r f a c e f i n d s worked w i t h s t e e l t o o l s f o r modern h a n d i c r a f t s
incliiding p s e u d o - t r a d i t i o n a l f i s h hooks. The p e a r l s h e l l was probably imported
from Nukuoro. Items G-I and L-N a r e p r e h i s t o r i c . K A l : I ( S p i t 2 ) , L ( S p i t 1 6 ) .
KA2:N(Layer 4 ) . KA4:H(Spit 1 5 ) , G ( S p i t 5),M(Spit 1 3 ) .
Sea Urchin Spine F i l e s
Four well-used f i l e s made from s l a t e pencil sea urchin
spines were recovered(see Figure 44). Three were found in the
Putau(KA3) s i t e dating to about 500 to 600 years BP, and the
fourth a t Tiroki(KAl) dating to about 200 years BP. The l a t t e r
i s hollow ground, and r a t h e r c h i s e l - l i k e ( # D ) . The other three
are q u i t e d i s t i n c t in being wedge-shaped, and these must have
functioned as abrasive implements for smoothing a f l a t surface.
Item #B has a wear facet on the very t i p , and i s not a reamer,
as might be thought from the i l l u s t r a t i o n . The reduced diameter
shown i s e n t i r e l y n a t u r a l . These f i l e s can not be i n t e r p r e t e d
-80-
as those for manufacturing one-piece fish hooks , where the reaming
operation necessary gives quite a different wear pattern. Instead,
they are most suited to finishing off some delicate flat
surface. A possibility is the flat areas on trolling lure
shanks, or arm rings. No comparable specimens have been found
on Nukuoro, where abrasive tools appear to have been made from
coral instead.
Pumice Abraders
Among the numerous small pieces of pumice recovered, quite
a few show evidence of having been used for grinding various
items. These are especially common in the Putau site(KA3)
from Spits 11 to 16, and Ngeiho-Hereu(KA4) in Spit 6. Similar
finds were also made on Nukuoro(Davidson,1971:71).
Worked Pearl Shell
Fragments of industrial pearl shell were extremely rare
in the excavations. Only six definite pieces of Pinatada
sp.
were recovered, in addition to the few artefact pieces. These
are illustrated in Figure 44.
©
eO
cm
Figure 46: Glass beads from the excavation at Muri-Harau(KA2). They are from
Layer 3(B), Layer 4 (A ,C,D,E,F) , and Layer 5(G). Some are dark blue(A,B),
some light blue(C,E,F,G), and one yellow(D). All are rather squat in crosssection, with the exception of #D, which is a long narrow bead.
-81Japanese and European Artefacts
P r a c t i c a l l y no modern c u l t u r a l material was recovered
from the surface layers of e i t h e r Tiroki(KAl), Putau(KA3),
or Ngeiho-Hereu(KA4); but such items were very common in the
excavation a t Muri-Harau(KA2). A s e l e c t i o n of items i s i l l u s t r a t e d
in Figures 45,46, and 47. The discrepancy between these
occurrences i s r a t h e r i n t e r e s t i n g , and the absence from deeper
l a y e r s even more so - i t serves to re-confirm a somewhat
b a t t e r e d f a i t h in s t r a t i g r a p h i c i n t e g r i t y for a Pacific
a r c h a e o l o g i s t . So commonly, Pacific archaeological s i t e s have
been severely and continuously disturbed by land crabs and
human digging and re-deposition a c t i v i t i e s t h a t close
s t r a t i g r a p h i c i n t e r p r e t a t i o n s can be q u i t e meaningless. The
absence of obvious c u l t u r a l mixing in the case of European
a r t e f a c t s on Kapingamarangi, shows t h a t t h i s problem i s not a
s i g n i f i c a n t factor on t h i s i s l a n d . European material i s found
in l a y e r s p r e c i s e l y where the land i s being b u i l t towards
F i g u r e 45: A s e l e c t i o n of European a r t e f a c t s from t h e excavation a t Muri-Harau
(KA2).A:brass c l a s p ( L a y e r 2).B:copper canoe patch(Layer 2 ) . C : b r a s s screw(Layer 2 ) .
D:iron n a i l ( L a y e r 1).E and F : b r a s s f i s h hooks(Layer 4 ) s e e a l s o Figure 39.
G:brass p o i n t ? f i s h hook(Layer 4 ) . H:brass c l a s p ( L a y e r 3 ) . I : ? c a s e i n comb
(Layer 3 ) . J : c o t t o n s t r i n g ( L a y e r 2 ) .
-82-
cms
Figure 47: European and Japanese artefacts recovered from the Muri-Harau(KA2)
excavation. A:American 5£ coin(dated 1964) from Layer 2.B:Japanese 10 sen
piece, (dated showa 17=1942) from Layer 2. C: Japanese 1 sen piece(dated
shove 2=1927) from Layer 3. D: metal button labelled "best ring edge" from
Layer 4.E:?shell button from Layer 3.F: button of Pi-nctada sp. from Layer 6.
-83-
Figure 48: Ceramic artefacts from the Japanese settlements on Nunakita
and Parakahi. There are parts of a food bowl(A), high voltage insulators
(B,I,J,K,L), rice bowls(E,F), plates(G), tea cups(H,M), and personal
containers (C,D).
the lagoon today. The absence in the other excavations must also
indicate that vertical land building on Touhou had practically
ceased by the protohistoric and historic period.At all these
excavations the loose surface rubble was cleaned away before
digging commenced.
Of the artefacts found, only a few need specific mention.
Of note, are two metal fish hooks, and a fragment of a third
{Figure 45, items E,F, and G ) . One of these is rather crude,
but the other more complete piece illustrates the typical
Polynesian flare for stylish fish hook curvature. The point #G
is from a larger hook/ and is very similar to examples from
Nukuoro(Davidson,1967:182,186) .
About 30 glass beads were found, some of them very tiny.
All but one were found in Layer 4. There are several types
evident, and three colour types are readily distinguished dark blue, lighter turquoise blue, and yellow. There are two
sizes - a larger, rather squat form, and a variable sized
smaller variety. The yellow bead is longer than the others.
-84Unfortunately, almost nothing is known of the characteristics
of early European trade items in the Pacific islands, and
this is an area where research should begin. It is to be
expected for instance that glassware, including beads, from
England, Spain and Germany would have distinguishable elemental
characteristics. The background work which would aid this
identification has not yet commenced. It is tempting to attribute
some of the Kapingamarangi glass beads to contact with Fernao
de Grijalvares in 1536; however, it must be noted that they
are quite different in character to glass beads found on
Taumako, in a context which almost certainly equates with
the visit there by Quiros in 1606. Necklaces of glass beads
of various colours and sizes were observed in 1910 on
Kapingamarangi(see Eilers,1934:119) . Glass beads have not been
found archaeologically on Nukuoro. In passing, it may be
mentioned that similar sized tiny glass beads of about 1mm
diameter have been found archaeologically on Banaba island
(Lampert,1965:19,23) to the east of Kapingamarangi.
-85-
KAPINGAMARANGI AND OTHER POLYNESIAN OUTLIERS
Linguistics
The Kapingamarangi language(KAP) is Polynesian in
character, and most commonly thought of as closest to that
spoken on Nukuoro(NUK), despite the fact that the two languages
are not mutually intelligible(Carroll,1965:193). Carroll's
study indicated a cognate percentage of 59% between these two
languages, although he noted that this could be inflated because
his KAP informant lived on Nukuoro and had used that language
for some time(ibid.:194).
Pawley has examined the relationship between the Polynesian
outlier languages in detail, and draws attention to the fact
that KAP and NUK "share a considerable number of morphophonemic innovations
and lexical items apart from other known languages
these shared features require either the assumption
of genetic sub-relationship, or very extensive borrowing"
(Pawley,1967:283-284) .
The question which linguists are unable to answer, of course,
is precisely when such borrowings took place - for instance,
could they be 19th century in origin, before the languages
were recorded in detail ? Pawley suggests at least three
Samoic Outlier(SO) subgroups (i)
Northwestern Outliers
KAP and NUK
(ii)
Central Outliers
Takuu(TAK),Luangiua(Ontong Java,LUA)
Sikaiana(SIK)
(iii) New Hebrides Outliers
West Futuna(WFU) and Mele-Fila(FIL)
He also raised the possibility the members of the first two
putative groups belong to a wider group with Ellice Islands(ECE)
dialects(1967:282).
Bayard's study suggests that Caroll's figures are
consistently inflated by about 12% (1976:61), and in his assessment
the nearest values to KAP have a somewhat different order,
as follows(after Bayard,1976:30):
Luanguia(LUA)
Takuu(TAK)
Rennell(REN)
Pileni(Taumako,PIL)
Tongic(TO)
Nukuoro(NUK)
Samoa(SAM)
44.3
4 3.7
4 3.7
42.0
37.1
35.7
35.5
-86Thus, in the case of the NUK-KAP cognate figure, Bayard's
findings are 23.3% lower than that obtained by Carroll.
Of the relationship between these two languages in particular,
Bayard comments:
"The two northern atolls seem dissimilar in their
correspondences despite a relatively high figure shared
between them. Even allowing for the possibility of
inflated percentages, KAP appears consistently higher
to the central atolls of TAK, LUA, and SIK than does NUK,
which shows high figures only to SAM, TON, and KAP."
(Bayard,1976:34).
Drift Voyaging
As pointed out above, Kapingamarangi is one of the
most isolated areas of land in the Pacific, and one which
seafarers continue to find difficult to make a landfall at.
Part of this difficulty is the apparently unpredictable nature
of the surrounding ocean currents; while these are generally
westward in direction, it is clear that they are eastward
at times too ! In view of this, it is extremely unlikely that
two-way navigation between this island and any other was ever
a significant factor influencing the cultural history of the
population.
Some drift voyages, however, probably arrived on the
island from time to time; indeed the original landfall must
have been just such a journey. In considering the founding
settlement of Polynesian outliers, Ward and others have
contributed a great deal of information on the possible role
of drift voyaging(Levison et al. ,1973; Ward et aZ.,1973),
and Bayard has recently reviewed this material(1976).
The relevant simulation voyaging assumed a point of
origin from one of five possible areas Tonga(five islands)
Wallis area(four islands)
Ellice area( two islands)
Tokelau(one island)
Samoa(two islands)
These experiments clearly showed that the northern outliers
had a higher possibility of contact from the Ellice area, with
the Wallis area a close second(Ward et
al.,1973:338).
-87For example, the Kapingamarangi figures are
Ellice
Wallis
Tokelau
56%
35%
9%
Those for Nukuoro, on the other hand are:
Ellice
Tokelau
Wallis
Samoa
36%
31%
18%
15%
One has to be very careful not to misinterpret this information.
Ward et al. have rightly warned that caution should be used in
interpreting their Table, as these figures indicate relative
orders, not exact probabilities(ibid.:337-8). In fact, when the
number of actual landings is taken into account, the contact
probabilities of these drift voyages to Kapingamarangi is
negligible(Ellice=0.9%, Futuna & Wallace=0.4%, remainder=zero);
moreover, the figures to Nukuoro are even less from these
places(see Levison et al.j1973:57).A total of 59 successful
landings were made to Kapingamarangi, compared with 4 79 to
Ontong Java, and this clearly shows the relative isolation of
Kapingamarangi atoll(it is not clear from Ward et al.,1973:
334
whether these landings are from the full 100,000 attempted
drift voyages, or a smaller number).
Despite this high percentage figure for Ellice Islands
cited above, it would be a serious error to assume that the
information on drift voyaging in any way points to an origin
of the Kapingamarangi people from this island area. It would
be interesting to have 'reverse' voyaging attempted from
Kapingamarangi, as this is bound to be much more enlightening
on likely sources of drift voyages to Kapingamarangi.
It is strongly suspected that if this were done, significant
contact with islands to the south would be indicated. As
Levison et aZ-.^point out in summary:
"our probabilities to the more distant Outliers,
such as Nukuria and Nukuoro[and presumably Kapingamarangi],
are negligible, but would probably be high or very high
if voyages were started from Ontong Java or Sikaiana"
(Levison et
al.3
1973:58).
-88Comparison with Nukuoro
The nearest neighbour to Kapingamarangi is Nukuoro,
some 215 km distant; the next closest land is also a Polynesian
outlier, Nuguria at 511 km(see Figure 1 ) . Despite the proximity
of Nukuoro, some interesting differences in prehistoric economy
and material culture have been revealed by the archaeological
research. It is worthwhile to summarise briefly some of the
more notable points of comparison between these two islands.
Fishing:
this was clearly a very important activity on
both islands at all periods. Detailed comparison must
await further analysis of the Nukuoro assemblage, now •
underway.
Shellfish:
on both islands, the role of shellfish appears
not to have been very important in the past; and Turbo sp.
is apparently equally predominant in the archaeological
sites.
Birds:
in neither island could birding be seen as a very
significant economic activity. Gallus
gallus
appears late
in the prehistoric sequence on Kapingamarangi, but has
not been found archaeologically on Nukuoro.
Pig: This is not an important aspect of the economy on
either island, and only appears in late archaeological
contexts, assumed to be historic in origin.
Dog: this is an important point of difference between
the two islands. Dog was never present on Kapingamarangi,
yet it occurs in early deposits on Nukuoro. It apparently
died out fairly early in the sequence there, and was
re-introduced in the historic period. The lack of dog on
Kapingamarangi is good evidence for isolation between the
two islands.
Rat:these are common in archaeological deposits on both
islands, but different species are involved - Rattus
exulans
on Kapingamarangi, and Rattus
rattus
cf. mansorius
on
Nukuoro. This difference is again compelling evidence of
isolation between the two islands, and also points to
contact between Nukuoro and islands further to the north
or west.
Turtle:
not important in either island.
Taro: the information here is not conclusive, but there
appears to be a notable difference. Although three species
of aroid are present on Nukuoro today, early accounts
mention only two - Coloaasia
esulenta(taro),
and another
which may have been either Alooasia
maarorrhiza
as stated
by Kubary(1900:108-9), or Cyvtosperma
ohamissonis,
as
implied by Eilers(1934:189). The name pura or puraka is
given by both Kubary and Eilers. It seems most likely that
Coloaasia
and Cyrtosperma
were present in pre-European
-89-
times, and that Alooasia
with its modern name ngaungau
is a more recent introduction. On Kapingamarangi,
By comparison, only Colocasi-a
esaulenta,
known as
tara,
is thought to have been present in the prehistoric period,
and the other two tare forms were apparently introduced
during the historic era, together with their Nukuoro
names. Thilenius(1903:17), on the other hand, claims that
not even the taro was present on Kapingamarangi before
the advent of the European;this point of view is also
put forward by Ray(1917:188).
Turmeric:
The cultivation of this plant and its associated
ritual was important in pre-European times on Nukuoro
(Kubary,1900:75). The plant was introduced onto
Kapingamarangi in the historic period.
Breadfruit:
this was almost certainly present on both
islands prior to European contact(unusual for atolls),
and bears the standard Polynesian name in both places.
On Kapingamarangi, a specialised preservation technique
occurs. This is not known on Nukuoro. The absence of the
paku-kuru
conserve on Nukuoro, again suggests isolation
between the two islands.
Shell
Adzes:
those made from Terebra
and Mitra
spp. are
as common as the ventral lip Tridacna
sp. form on
Nukuoro, and inner hinge Tridacna
sp. adzes are about
one-third as numerous as either. On Kapingamarangi, the
ventral lip Tridacna
sp. adze almost totally dominates.
Hinge adzes are rare, and gastropod adzes are either absent
or insignificant in quantity. This difference is not
related to any environmental factors, and isolated
development is again indicated. It is also notable that
turtle bone adzes are common in the Nukuoro archaeological
sites(Davidson,1971:76), but absent on Kapingamarangi.
One-Piece
Fish hooks:
on Kapingamarangi, the near absence
of Pinctada
sp. meant that non-durable forms dominated,
such as those made from coconut shell and turtle shell.
The fish hook forms are simple, but distinct from those
on Nukuoro. This probably reflects the influence of
material constraints, rather than lack of contact as such.
On Nukuoro, on the other hand, there is an abundance of
pearl shell by comparison, and many hooks were made of
it in several distinct forms. Metal hooks, based on
prehistoric styles, were made on both islands in the
early historic period.
Trolling
Lure Shanks:
these are present in small numbers
in both islands. On Nukuoro, several types were made
which are typical of many in the Pacific. A unique form
was developed on Kapingamarangi, again indicative of
isolation.
-90Abrasive
Files:
on Nukuoro these are made of coral; and
on Kapingamarangi they are of sea urchin spines.
Arm Rings:
the triangular-sectioned Tridaona
sp. arm
ring is present in early contexts on both islands. Later
examples are known on Nukuoro made from a large bivalve,
and from Conus sp. on Kapingamarangi.
Coconut Grater
Heads: the use of unfashioned bivalves for
this purpose appears to have been common throughout the
sequence on Nukuoro, yet rare or absent on Kapingamarangi,
Specially made heads of Pinotada
sp. appear relatively
late on Nukuoro, and on Kapingamarangi specially made
heads of several kinds of shell are present throughout
the prehistoric sequence.
Loom Belt Weaving:
this was present on both islands; on
Nukuoro it is essentially a female activity, as it is
generally in the Caroline Islands. On Kapingamarangi,
however, it is a male prerogative, in common with almost
all Polynesian outliers in Melanesia(Riesenberg and
Gayton,1952:350). Moreover, hibiscus fibre was used
on Kapingamarangi, and banana fibre on Nukuoro(Buck,
1950:148) .
Tattooing:
This was practiced on Nukuoro, on women
especially(Kubary,1900:80-81); but was almost certainly
absent on Kapingamarangi.
The points of similarity mentioned above probably reflect the
general family resemblance which one could point to for any
pair of Polynesian islands in the Pacific, regardless of
how distant from each other. The differences observed, however,
are so great as to suggest substantial if not total isolation
from each other, once primary settlement had been achieved.
-91-
CONCLUSIONS
Land H i s t o r y of t h e
Atoll
Kapingamarangi a t o l l h a s n o t been a s t a t i c s e r i e s of i s l e t s
o v e r i t s r e c e n t g e o l o g i c a l h i s t o r y . On s u c h a l o w l y i n g s p e c k
of l a n d , t h e s e a r a g e s a c o n t i n u o u s war of a t t r i t i o n o v e r t h e
b a r e l y c o n s o l i d a t e d c o r a l g r a v e l s w h i c h make up e a c h i s l e t . T h e s e
g r a v e l s a r e m e r e l y s i t t i n g on t h e l e e w a r d s i d e of t h e main c o r a l
r e e f , and sudden s t o r m s can wreak havoc w i t h t h e i r s t a b i l i t y .
T h i s i s e v i d e n t f r o m b o t h t h e known r e c e n t h i s t o r y o f t h e i s l a n d ,
and from k n o w l e d g e o f l a n d - b u i l d i n g p r o c e s s e s p o s s e s s e d by t h e
a t o l l ' s inhabitants. Matukerekere, the tiny i s l e t to the
s o u t h , was d e v a s t a t e d by a s t o r m e s t i m a t e d t o h a v e t a k e n p l a c e
i n 1 8 5 8 , a n d i s now s l o w l y r e - e s t a b l i s h i n g i t s e l f . T h e l o n g
i s l e t of Hare h a s t h r e e l a n d d i s t r i c t s on i t s s o u t h e r n p a r t ,
which were s e p a r a t e i s l e t s u n t i l r e c e n t l y . Only one of t h e s e
i s l e t s h a d b e e n c o l o n i s e d b y v e g e t a t i o n i n 1 9 4 2 , t h e Luawa
d i s t r i c t ( s e e F i g u r e 4 9 ) . I t i s l i t t l e wonder t h a t t h e P o l y n e s i a n
h a s f o r i t s common m e a n i n g ' u n c u t ' ,
w o r d f o r i s l a n d o f motu,
' n o t - d i v i d e d ' o n K a p i n g a m a r a n g i - a double-entendre
nicely
d e s c r i b i n g t h e t e n u o u s hold t h e l a n d h a s o v e r t h e s e a on t h i s a t o l l .
Figure 49: The s o u t h e r n t i p of Hare i s l a n d a s i t was i n 1942. Three small i s l e t s
can be seen, one of which i s i n t h e p r o c e s s of j o i n i n g onto t h e main block of
land.By 1950, f u r t h e r land j o i n i n g had taken p l a c e ( E m o r y , 1 9 6 5 : 4 ) , and by 1980,
only t h e small i s l e t t o t h e r i g h t was s t i l l s e p a r a t e . This p r o c e s s of p e r i o d i c
i s l e t r e c l a m a t i o n and s u b d i v i s i o n i s b e l i e v e d t o have been underway f o r a t
l e a s t as long as people have l i v e d on Kapingamarangi.
-92The people on Kapingamarangi are thoroughly aware of this
power the sea has to take away their land, and to build it up,
and they take advantage of it. They engage in some very clever
examples of land building of their own, in which they harness
marine power to do the work. By judicious wall-building, the
sea can be encouraged to make new land, by subtle alterations
in the currents this causes. Even apparently insignificant
lines of rocks placed on the reef can have a substantial
land-building effect. Knowledge of this kind of process could
easily be acquired by observing the effects of making fish
traps with rocks in the shallows. By consolidating newly
formed land with more substantial walls on its periphery, a
somewhat more permanent 'islet' can result.
Apart from Touhou islet, almost all of the present
land in Kapingamarangi is very low to the sea, and no
archaeological deposits could be found on other islets which
indicate any lengthy stratigraphic history. A distinct possibility
therefore is that the land has come and gone many times at
different stabilizing points on the reef over millennia;
and that the present configuration of islets is merely one
chapter in the atoll's complex geomorphological history.
Although uncertain, it is strongly suspected there are no
culturally deposited sediments anywhere on Kapingamarangi•atoll
which are older than those on Touhou islet. Yet, when cultural
debris was first deposited on Touhou, about 800 to 1000 years
ago, this land was barely above sea level, and could not
have sustained domestic habitation in the islet's present
form. In short, the people leaving this first cultural
evidence must have been in residence on some other islet in the
atoll chain. Just how much longer the full prehistoric sequence
of occupation is for Kapingamarangi will never be known, because
the islets these pre-Touhou people lived on have, in our view,
subsequently been stripped of cultural soils. Some early massive
storm which could have had this effect, may well have been the
deciding factor encouraging people to take up residence on this
slowly consolidating patch of sand that was to become Touhou
in due course. Over the next 500 years, the people imported
massive quantities of coral gravel onto this islet, and built
it up to its present-day height of 4m above the sea. An added
advantage of this relatively high artificial island is that
it would serve to strengthen the fresh water reservoir in its
basal sediments. This would act as a useful buffer for the
water supply in times of drought. Kapingamarangi is in an
area of the Pacific with unreliable rainfall, and any means
to lengthen the available buffer would have been worthwhile.
It is doubtful though that the people realised this advantage
that island building bestows. It is notable that in the course
of the formation of Touhou islet, the density of fish bone
material increases in the sites five-fold - this must surely
reflect in part the growth of the population in residence on
the islet. The early people also took the precaution of
building a large sea wall on the ocean side of the islet,
and this is rapidly replaced each time it gets washed away
by the inevitable periodic storm or exceptional tide.
-93Thus characterised, the artificial islet of Touhou is seen
as an appropriate adaptation to life on an atoll where the sea
constantly challenges the permanence of land.
Origins of the Founding Population
An intriguing question is where precisely did the earliest
Kapingamarangi people come from ? Unfortunately, it is not
possible to be certain on this point. In the future, the
best possibility for establishing a satisfactory answer to it
may be by acquiring additional knowledge from human skeletal
material, not only from Kapingamarangi, but from other areas
in Micronesia, Polynesian outliers, and especially in Melanesia.
This type of information is seldom incorporated in discussions
of the prehistory of individual islands, and this is a little
like observing the stage set of a theatre without the players it leads to imperfect knowledge of the play. At this stage
we can be quite confident that the founding group came from
an area already inhabited by Polynesians - this is evident
from the scant skeletal material recovered, and of course
the language spoken at the close of the prehistoric period.
However, we do not know exactly how long people have been on
Kapingamarangi, nor very much about their early material
culture. It appears that the island has been inhabited for
at least 700 to 1000 years. So far as is known, Nukuoro to the
north was inhabited for a similar length of time, but here it
appears that a reasonable sized islet existed when occupation
first began. It is therefore unlikely the Nukuoro prehistory
extends back much beyond what is presently known. The same
cannot be said for Kapingamarangi, as it has been argued above
that there must have been a population in residence elsewhere
in the atoll before the earliest sediments on Touhou. Thus, present
evidence favours Kapingamarangi being settled before Nukuoro,
and this cuts out one possible island for the source of the
founding group on Kapingamarangi. Of course, these arguments
about landform history depend on the assumption of land-sea
stability over the last thousand years or so. However, these
atolls are not far from the unstable west Pacific zone of
tectonic movements and volcanism, and earthquakes are not
unknown on Kapingamarangi(Wiens,1956:12). Assuming tectonic
stability, present thinking would be against occupation of
these northern atolls earlier than about 1500 years at the
outside. The suggested independant settlement of Kapingamarangi
and Nukuoro is more positively indicated by the differences
outlined above in material culture and economy between the
two islands. Of special importance in this respect is the
different species of rat on the two islands.
Traditionally, the people on Kapingamarangi trace their
descent from a man called Utamatua and his wife Roua, who came
from an island called Tamana, although another island Nikuhetau
is also mentioned in connection with the pair(see Eliers,1934:
126ff; Emory,1965:28,35). Tamana is described as one of three
islands, the others being Pakihi and Pararu(Elbert,1949:240).
There is a Tamana island in the Kiripati group, and Nukufetau
-94is known in the Ellice group. The identity of the other two
islands is unknown. Nukufetu and Kapingamarangi are both
mentioned in origin myths on Nukuoro(Eliers,1934:129); and
in Ontong Java, Kapingamarangi[Makarama] is mentioned in origin
tales(Thilenius, 1903:31). These traditions surely indicate
knowledge of other islands surrounding Kapingamarangi; but they
should not be interpreted literally on the question of origins.
It is more likely they contain a measure of information gained
from the arrival of rare castaways on the atoll from elsewhere.
To interpret the Utamatua story literally one must take into
account that the story is primarily about the creation of
Touhou islet(Elbert,1949:240), and that a man called Korae
was already living on Turuaimu, when Utamatua arrived on
Kapingamarangi. One may well ask who was this earlier inhabitant
Korae, and where did he originate ? - rather than pay too much
attention to the origin of the interloper, Utamatua. This is
probably a good example of what is so frequently found in the
analysis of similar myths amongst the New Zealand Maori; their
role is essentially a charter for modern land ownership, rather
than an accurate record of history, which can be interpreted
literally.
Probably the soundest conclusion that can be reached
on present knowledge is that the first group on Kapingamarangi
arrived from somewhere to the south, probably from another
Polynesian outlier in Melanesia. It can be noted in passing that
artificial island building, such as that involved at Touhou,
is not uncommon on the coastal areas of Malaita, and of course
on Tauamko, a Polynesian outlier.On the basis of the linguistic
evidence, Bayard reaches the conclusion of "primary settlement...
from TAK,NKM, NKR, or SIK"(1976:53), and this view is not at
all at variance with the archaeological evidence. What is
somewhat more questionable though, is the view of significant
secondary contact with Nukuoro, Kiripati, and the Ellice
Islands in the pre-European era(ibid.) In our view, the
archaeological evidence indicates considerable isolation on
Kapingamarangi once settlement had been established. Contacts
with the outside world when they occurred do not seem to have
made a significant impact on the course of Kapingamarangi
culture-history. If the linguistic evidence demands a measure
of borrowing between Nukuoro and Kapingamarangi, then this may
well have followed contact with Europeans, and the resulting
increased mobility of Polynesians, during the 19th century.
A probable exception to the long history of isolation
on Kapingamarangi concerns the manufacture of a specially
wrapped food preserve on the island. This has been described
in great detail by Eilers(1934:72),Buck(1950:35ff), and
Emory(1965:134). Eilers maintains that this conserve, to bako ,
was made with pastes of both breadfruit and pandanus, and
draws attention to the similar pandandus conserves of Nauru(1934:
72; see also Hambruch,1914:113ff). Buck also notes the
similarity of the carefully tied roll of the breadfruit
conserve, paku kuru,with
rolls of pandanus from both Kiripati
and the Marshall islands. However, according to his informants,
pandanus was never treated in this way on Kapingamarangi(Buck,
-951950:30-31). This type of conserve is unknown on Nukuoro, as it
is on all other Polynesian outliers - how did it come to be
on Kapingamarangi ? This may well indicate an instance of
secondary contact with an island to the east in the Nauru or
Kiripati or Marshalls area. The possibility of some prehistoric
contact with this area is further supported by the type of
roof thatching on Kapingamarangi. People on the island claim
that this .was formerly of coconut leaves, and that pandanus
thatching was introduced by a castaway couple from Tarawa in
the Kiripati group(Buck,1950:66;Emory,1965:52). On linguistic
grounds alone, an origin for the Kapingamarangi
population from the Kiripati - Marshalls area is out of the
question. These two instances of close cultural similarity
must be examples of borrowing in the pre-European period,
suggesting occasional drift voyages from the outside world
to the east.
The other difficulty with the linguistic evidence is
that archaeological findings now confirm that a number of
Polynesian outliers(Anuta, Tikopia and Taumako at least),
have been occupied for up to 3000 years, while the language
time depth appears only to be about 1000 years. For Taumako,
at least, there is no reason to propose that the earliest
people were not from Polynesia itself. Thus, relationships
established on the basis of languages spoken in the 19th
century, may merely document the last chapter in a very complex
story; and while linguistics confirms the 'backwash' theory
of Polynesian outliers, it is poorly equipped theoretically
to cope with what now seems likely - multiple backwashes, perhaps
from several areas. On linguistic grounds, the Ellice Islands
appears to be a prime candidate as the direct source of a
number of Polynesian outliers. It is very important in the
future to see whether the Ellice Islands has an occupation
time depth of 3000 years, and until this is established it
cannot be considered as the source for the earliest populations
on some of the outliers. So far as Kapingamarangi is concerned,
this problem over relative time depths does not arise - it is
unlikely that the full prehistoric sequence would be more
than 1200 to 1500 years. The main difficulty in deriving the
earliest Kapingamarangi population from the Ellice group is
the lack of close resemblance between material culture in the
two regions at European contact. A long list of items could
be drawn up, particularly in the realm of non-durable material,
which simply do not occur on Kapingamarangi. This includes
tattooing, reef sandals, turtle bone adzes, ray skin rasps,
a special kind of rat trap, certain kinds of small hand nets,
the teka dart, drums, and so on(see Hedley,1897 ,-Kennedy,1931;
Koch,1961). Their absence on Kapingamarangi is made more
significant by the fact that many of these items are present
on other Polynesian outliers to the south, such as Taumako.
A further point concerns the remarkable fishing technology
in the Ellice Islands devoted to catching palu,
the ruvettus
fish(Kennedy, 1931:12ff). It is difficult to believe that
the masterly fishermen of Kapingamarangi, once they knew of
such a valuable fish for the taking, would not exploit
-96the resource to some degree. This fish has a circumtropical
distribution, even occurring in the Mediterranean(Gosline and
Brock,1960:253). It was not collected by the fishermen assisting
with the comparative collection during the expedition, and this
is bound to reflect a lack of knowledge of its habits, rather
than a lack of presence of the fish. This type of fish is only
caught in depths of 150 to 450m off the sheer face of the
reef(Kennedy,1931:29). Bones of these fish are readily identified,
although they may well have a lower survival rate than those
of other species. This is because of the oily nature of the
bone, rather like that of tuna and barracouta, though more so.
The bones can even be eaten with the flesh(see Titcomb,1972:159).
Lack of survival archaeologically cannot be the full story,
however, as less durable cartilage from elasmobranchs was common
in the sites. We can be confident therefore that the Kapingamarangi
fishermen do not now, and never did, know how to catch this
valuable fish. Although it has yet to be established just
when the people in the Ellice Islands learned this technology,
the lack of it on Kapingamarangi does seem significant.
This discrepancy is pointed out with some reservations,
however, because it appears that ruvettus fishing was known
on Takuu, Nuguria, Ontong Java, Nukumanu, and Sikaiana.
This conclusion is based on the presence of the fairly
distinctive large wooden hook with a detachable strongly incurved
point(Anell,1955:231). Large, but dissimilar hooks were made
on both Kapingamarangi and Nukuoro(ibid.:230 ;Buck,1950:234;
Beasley,1928:101), but in both places were said to be used for
catching sharks, although these would also function for other
large deep water fish. It is not known how long this fishing
technology was present on these Polynesian islands in
Melanesia, and it is not universal amongst outliers by any
means. The Taumako people, for example, had only a poorly
developed fishing technology, and ethnographic specimens of the
ruvettus hook are unknown from this area.
On the whole, the evidence of ruvettus fish is difficult
to understand fully at present. It would probably be wise to
reserve judgement in the meantime, and await better documentation
directly from bones of the fish archaeologically.
At this stage it would be a mistake to attempt to be
too precise on the question of the origins of the founding
settlers of Kapingamarangi. Very little is known of the
prehistory of the closest Polynesian outliers in Melanesia,
but there is every sign of complexity. For instance, although
traditions on Ontong Java identify Kapingamarangi as their
original homeland, a close biological relationship between these
two populations seems very unlikely. This has been vigorously
opposed by Shapiro (1933:271), and there is no reason to update
this view. Similarly, close biological affinity between
Kapingamarangi and Nukumanu has been equally strongly opposed
(Schlaginhaufen,1929:282), although in this case the relationship
between those people studied, and the earlier inhabitants of
Nukumanu may be questioned. The only Polynesian outlier where
it has been possible to examine early skeletal material in
-97any detail is Taumako in the Solomons. Biological affinity
between Kapingamarangi and this island could certainly be
proposed; however, the affinity is probably due to general
family resemblance between Polynesians, rather than direct
contact in this case. There are several artefact forms on
Taumako which have been important to the population there
for nearly 3000 years(see Leach and Davidson,n.d.); it seems
unlikely that these would not occur on Kapingamarangi, had
there been direct contact from this area. In addition, the
severe incidence of yaws on Taumako makes it quite likely
that settlement from this island would have involved
transportation of the disease.
In summary, there is every reason to believe that the
populations on Nukuoro and Kapingamarangi were seeded by
immigrants from different areas. Present evidence favours
a southern source for Kapingamarangi, although it is not
possible to be more explicit than this. It is evident that
after initial settlement, occasional contact occurred with
islands to the east, perhaps with Nauru or Banaba.
The Prehistoric Fishermen of Kapingamarangi
Pitting one's wits against the sea and the fish it
contains is of foremost importance to all men on Kapingamarangi.
Endless hours are spent discussing the whys and wherefores
of different kinds of fish, their habits, and how to catch them.
In this respect, the Kapingamarangi fisherman is little
different to his counterpart in modern European society - however,
for skill in fishing and knowledge of fish habits, the men on
Kapingamarangi are second to none. It is hardly surprising
that in Kolonia on Ponape, the men of the Kapingamarangi
village of Porakiet have established themselves as the mainstay
of the fish market, and their prowess as fishermen is supreme.
Above all others, the type of fish which is seen as the most
worthy adversary is the pelagic game fish, which is frequently
caught by trolling a lure in the surface of the sea over
deep water. This type of fish is a voracious carnivore, feeding
on small fish which it drives to the surface, and then seizes
from beneath. Frequently, fish of this kind work in packs
to drive a shoal of smaller fish together. Such packs are
easily recognised either from the birds which follow them
about, or by the surface turbulence caused by small fish
darting to and fro in their attempts to escape. If the fisherman
manages to get into such a position with a lure, an abundant
catch is assured, because the game fish in their feeding
frenzy will simply attack anything moving in the water.
In addition, packs of game fish can be encouraged together
from a distance to some extent, by trolling a lure through the
water; its action disturbs the water, and emulates a small
struggling fish at the surface, and it is therefore an
irresistible attraction.
The species of fish which can be taken by this method
of fishing, vary from one part of the Pacific to another.
On Kapingamarangi, the main ones are as follows:
-98Agrioposphyraena
barracuda Barracouta
Aaanthoaybium 'solandri
Wahoo, Jack Mackerel
Neothunnus macropterus
Yellowfin tuna
Katsuwonus pelamis
Skipjack,striped tuna
takua,
and te atu3
These fish are known as te ono, te mala,
names which are almost universal in the Pacific, and which
reconstruct at least to proto-Polynesian(see Appendix 1 ) .
In the Kapingamarangi language, special terms exist which
relate to small variations in size and characteristics, clearly
showing the central importance of these fish. It is hardly
surprising that when fishing for our comparative osteological
collection began on Kapingamarangi, the first two species of
over 100 eventually collected were takua and te mala.
As far as the modern Kapingamarangi fisherman is concerned,
to find out anything useful or interesting about the prehistory
of fishing on the atoll would be to illustrate the varying
prowess of his ancestors in catching these types of fish - the
evidence shows that while much dreaming of these fish may
have taken place in the past, far more were probably caught
in the men's house than ever in a canoe !
This discrepancy between the spectacular bonito or
tuna fishing activities of people in a number of areas of the
Pacific in the historic period, and the pattern of catches
revealed by the archaeological record is a considerable mystery
at present. Even in the temperate waters of New Zealand,
where bonito or tuna fishing was not really feasible, the same
discrepancy can be seen in the counterpart fish - the kahawai,
Arripis
trutta.
Special lures for these fish are amongst the
most common items in 19th century ethnographic specimens
collected from the New Zealand Maori; yet both the hook and
bones of the fish are extremely rare in archaeological sites.
The same phenomenon is true for Kapingamarangi. Of all the
fish which could have been caught by trolling with lures,
only five in every 100 fish actually caught, belonged to this
type in the past. It would be interesting to know just how
frequently these fish are caught today; our suspicion is that
the true situation is more like the Pygmy elephant hunters
in Africa - there is much talk of elephants, and there is a
specialised vocabulary about elephants, but the diet largely
consists of items other than elephants. It would be a mistake
therefore to hurriedly seize upon this dearth of game fish in
the archaeological record, and argue that the prehistoric
fishermen of Kapingamarangi were far less skilled than their
modern descendants.
When Peter Buck visited the island in 1947, one of his
ambitions was to go "on a fishing expedition the purpose of
which was to view the method of line fishing for tuna"(Buck,1950:
219); but after several days away from Touhou, during which
-99many fish other than tuna were caught, he failed to observe
this activity at first hand. He commented that "fish seemed
to be scarce"(ibid.:248), and was told that "the day was wrong
(huaaitu)
because it was too bright and the fish were not
swimming near the bottom"(ibid.:24 9 ) ; eventually rather sadly,
they paddled through the reef channel and the four miles back
to Touhou. He adds "we had another demonstration of the value
of direct observation"(ibid.:249), which barely disguises the
disappointment he must have felt. Buck concludes that "bonito
fishing by trolling with a lure appears to be on the wane on
Kapingamarangi"(ibid.:237).
On the contrary, this experience could be viewed in
another way altogether. Catching pelagic fish is largely a
matter of seizing the opportunity when and if it arises. These
fish are rather unpredictable in their occurrence, and travel
far and wide. They can easily be caught whenever they are
actually present, either outside the reef or inside, but one
must be close at hand or actually in position when the opportunity
arises. To go out specifically to catch only these fish is to
invite an empty stomach in the evening. In point of fact,
as every modern European fisherman knows, when fishing for the
common and mundane fish, it pays to have a lure either
permanently trolling along behind, or immediately to hand,
in order to catch the pelagic fish when they turn up. When a
flock of birds is seen, one hurriedly pulls up the baited
lines, and dashes over to the vicinity of the birds, to try
and catch some of the pack. More often than not, the pack
disappears as quickly as it made its presence known, and
baited lines are dropped again to carry on the main, but less
spectacular fishing activity. Thus, the catch of even the most
expert fisherman, will largely be composed of something other
than pelagic fish. This is exactly what Buck observed, and
exactly what the archaeological record demonstrates.
What then can be seen about the main body of the catch
during the prehistoric period ? The most interesting aspect
of this is the abundance of parrotfish in the earliest levels
and the relative absence of members of the cod and groper family.
In the troical Pacific area the Scarids or parrotfish fill a
similar role to that of Labrid fish in more temperate waters
- that is, they are abundant and easy to catch. For this very
reason, they are not especially sought after, and in fact are
considered inconsequential or even despised by the expert
fisherman. It is ironical, but hardly surprising that Labrids
formed the mainstay of prehistoric fish diet amongst many
communities in early New Zealand, and Scarids form the basis
of fish diet of many people in tropical Oceania. It is
interesting that in the case of Labrids, some lines of evidence
suggest that these fish offer security in times of economic
or environmental stress(Leach and Anderson,1979a), and are
therefore more abundant during periods of hardship. A similar
hypothesis might be advanced for Scarids too. Seen in this way,
the early archaeological evidence on Kapingamarangi might
show somewhat less environmental familiarity than in later
periods. Scarids are primarily herbivores, and as such do not
-100normally take a baited hook. They are therefore caught primarily
by netting, and many can be taken in this way in relatively
shallow water near coral thickets. Catching this type of fish
is a 'safe' activity, guaranteed to succeed. On the other hand,
baited line fishing, such as is appropriate to the cod and
groper family, requires somewhat more initiative and skill
in deeper water, is less secure, more adventuresome, and not
always guaranteed to produce a lot of fish. A suitable index
therefore of the real skill of prehistoric fishermen is not
so much how frequently pelagic fish are caught, but the relative
quantities of cod and groper as compared with parrotfish.The
relevant figures for the different periods on Kapingamarangi
are given in Table 11 below:
TABLE 11: TRENDS IN THE RATIO OF PARROTFISH TO COD AND GROPER
IN KAPINGAMARANGI
Period
Cod & Groper
Parrotfish
Ratio
0-100 BP
23.26%
20.93%
1.11
100-300 BP
19.75%
18.81%
1..05
300-700 BP
15.99%
17.27%
0.93
7.69%
23.08%
0.33
700-?1000 BP
This simple ratio shows a uniform rise in the relative
importance of baited line fishing over the last millennium,
and must indicate a dramatic improvement in fishing skills
and technology over this period. With the near absence of
suitable shell for making one-piece bait hooks, this increasing
abundance of fish caught on hooks made from less durable
materials such as coconut shell, is a striking testimony of
the prowess and enterprise of the prehistoric fishermen of
Kapingamarangi.
-101-
APPENDIX 1
MARINE SPECIES IN KAPINGAMARANGI ORTHOGRAPHY,
WITH SOME PROTO-POLYNESlAN EQUIVALENTS
D.T.Bayard
For ease of comparison, terms in the text have been
rendered in a standard "Polynesian" orthography. However, the
orthography used on Kapingamarangi and in Lieber and Dikepa's
KAPINGAMARANGI LEXICON recognises the distinction made between
long (aspirated) and short (unaspirated) consonants in the KAP
language (Lieber and Dikepa, 1974: xvii-xviii). This distinction
has arisen largely as a result of the disappearance of a medial
vowel between two identical consonants (*C1VC1 > ClCl); hence
PPN *totolu
becomes ttolu
<tolu> in contrast with
tolu
<dolu> (Pawley, 1967:286). The following long/short pairs of
phonemes are distinguished in the KAP orthography: p/b (<PPN*p);
mm/m (*m) ; ww/w (*w) ; t/d (*t) ; nn/n (*n) ; k/g (*k) ; nng/ng
(*ng); hh/h (*f/*s); 11/1 (*l/*r). Where possible, ProtoPolynesian equivalents have been added (from Walsh and Biggs,
1966; Biggs, 1979; Pawley and Green,1971); however,
exhaustive search and investigation of irregular reflexes
would doubtless increase the number of proto-forms presented
here.
The family numbers given below follow the order in
Munro (1967: vii-xv). The fish and shellfish species, together
with the PN names, are those collected and identified by Leach
and Ward during their expedition.
FISH SPECIES
Family
Species
Car charhinus
spallanzani
Triaenodon
apicalis
Amphotistius
kuhli
Megalops
cypvinoides
Chanos ahanos
Gymnothorax undulatus
Tylosurus
indicus
Rhyneorhamphus georgi
Arnoglossus
cf.intermedins
Ostiahthys murdjan
Ostichthys
cf. adustus
Holoaentrus cf.
tiereoides
Holocentrus ruber
65 Sphyraenidae
Sphyraena
jello
Agrioposphyraena barracuda
66 Mugilidae
Valamugil
seheli
72 Scombridae
Grammatorcynus
bicarinatus
73 Acanthocybiidae Acanthocybium
solandri
75 Thunnidae
Neothunnus macropterus
76 Katsuwonidae
Katsuwonus pelamis
88 Carangidae
Caranx melampygus
Carangoides cf.laticoudis
Caranx cf. melampygus
Caranx
sexfasciatus
Caranx cf.
ignobilis
Carangoides oblongus
Elegatis
cf.
bipinnulatus
97 Epinephelidae
Cephalopholis
miniatus
Plectropoma maculatum
Cephalopholis
oyanostigma
Cephalopholis sp.
Epinephelus cf. tauvina
Aethaloperca rogaa
4
5
7
15
21
42
47
48
53
57
Carcharhinidae
Triakidae
Dasyatidae
Megalopidae
Chanidae
Muraenidae
Belonidae
Hemirhamphidae
Bothidae
Holocentridae
"PN" Name
KAP Name
hokoulu
mokolewe
te hai
?
te nga
matakalati
te kiha
te aku moana
te iha
te paipai
malau pungu
malaupunga luli
malau
te ta
tapatu
te ono
kanae
ika-pou
te mala
takua
te_atu
alala
laloale motomoto
(no name)
te kata
te kata lehu
laloale
te kina
mataili
?
tai-awa
matapaku
(no name)_
ngatala talo
taea uli
hogoulu
mogo-lewe
di hai
di ngaa
mada-galadi
di gihaa
di agu moana
di iha
di baibai
malau-pungu
malaupunga luuli
malau
di daa
dabaduu
di ono
ganae
iga-bou
di mala
dagua
di adu
(gada-)alaala
laloale-modomodo
gada
di gada-llehu
laloale
di gina
madaaili
daiawa
madaabagu
ngadala-daalo
daaea-uli
PPN Term
*fai
*ise
*malau
*malau
*malau
i
*'ono
*kanahe
*maratea?
*takua
* 'atu
H
O
M
97 Epinephelidae
Epinephelus
Epinephelus
Promiarops
Epinephelus
Epinephelus
106 Lutjanidae
107
108
110
114
Aphareidae
Caesiodidae
Nemipteridae
Lethrinidae
damelii
fuscoguttatus
lanoeolatus
fuscoguttatus
sp.
Plectvopoma maoulatum
Pleotropoma maoulatum subsp,
Lutjanus mono stigma
Lutjanus
semioinatus
Maoolor sp.(not M.niger)
Lutjanus cf . aoatesi
Macolor niger
Lutjanus
janthinuropterus
Aprion
viresoens
Lutjanus
sebae
Lutjanus
rivulatus
Aphareus cf.
rutilans
Caesio
coerulaureus
Monotaxis
granocutis
Lethrinus
fletus
117 Mullidae
Parupeneus
124 Kyphosidae
125 Platacidae
Pavupeneus macronemus
Kyphosus cf.
vaigiensis
Platax
orbicularis
129 Chaetodontidae
130 Pomacanthidae
138 Coridae
140 Scaridae
Chaetodon
cf .
cf.
janseni
semeion
Heniochus
varius
Pygoplites
diacanthus
Epibulus
insidiator
Ch.tvilobatus,
Ch.undulatus
Thalassoma lunare
Cheilinus
fasciatus
Eemigymnus melapterus
Callyodon
cf.
niger
Bolbometopon
bicolov
Cetossarus
putchellus
Bolbometopon
muraticus
Scarops
jordani
manukuou
pulungame
?
te mohia
ngatala te
ngatala-moko-te
te tonu
tonu matahale
tangau ka
tangau pulapula
te kuku
te tata
?
te hina
tangau poto[ina?]
te utu
te taea
tangau poto[ina?]
matamata
te uli
te mu
te muka
te kala
te matuailau
ngatala akau
kanikani hatu
tihitihi
tihitihi
talehala
?
?
te lipai
te lali
te lali
talinga mea
toliki
te huhu
paheni uli
paheni
melenge
te huhu
manu-guou
bulu-nga-mee
di mohia
ngadala-dee
ngadala-mogo-dee
di donu
donu-madaahale
dangau gaa
dangau bulabula
di gugu
di dada
di hina
dangau bodoina
di udu
di daaea
dangau-bodoina
madamada
di uli
di muu
di muu-gaa
di gala
?
di madu-ai-lau •
ngadala-agau
ganigani-hadu
dihidihi
dihidihi
dalehala
di liba-i
di lali
di lali
dalinga-mea
dooligi
di huhu
baaheni-uli
baaheni
melenge
di huhu
*tonu
*tonu
* 'utukao?
i
*huli
*muu
*muu ?
*kaloama
*matu
*tifitifi
*tifitifi
*'ufu
*'ufu
£
OJ
'
140 Scaridae
157 Siganidae
159 Acanthuridae
169 Scorpaenidae
175 Diodontidae
17 8 Tetrodontidae
180 Balistidae
182 Ostraciidae
Chlorurus
mic rorhinos
Hipposaarus
I ongiceps
latus
Siganus
cf. v ermicu
Siganus
rostr atus
Siganus
ahrys ospilos
Naso sp.
Naso
annulatu
Aaanthurus
cf . gahhm
Cyphomyoter
v lamingi
Aaanthurus
cf . gahhm
Naso
uniaorni s
Naso cf. brev irostris
Aaanthurus
ol ivaaeus
Calliaanthus
cf.
lituratus
Naso
annulatu s
Aaanthurus
cf . gahhm
Aaanthurus
cf .
pyroferus
Saorpaena
car dinalis
Pterois
volit ans
Saorpaena
car dinalis
Diodon cf. hy strix
Arothron
hisp idus
Arothron
stel latus
Pseudobaliste
s
flavimarginatus
Hemibalistes
chrysopterus
Meliahthys
bu niva
Ostraaion
tub eroulatum
palakia
ulahi
pongongo
paua
pongongo uli
lapia
te imu-alo
hupaka
te kelu
palangi
kelu akau
te ika hepu
te lipa ta
hukumea
?
imu-walu
te lipa kono ehe
te lipa
te nohu
te hauihau
(no name)
toutu
te wali
te ete
?
te imu
te humu hapula
humu hakapekepeke
tekemomo
balagia
ulahi
bongongo
baua
bongongo-uli
labiaa
di imu-alo
hubaga
di gelu
balangi
gelu-agau
di iga hebu
di liba-daa
hugu-mea
imu-walu
di liba-gono-ehe
di liba
di nohu
di hauihau
te keke
te hono
di gee
di hono
*'ulafi
*palangi
*nofu
doudu
di wali
di ede
di imu
di humu habula
*sumu
humu hagabegebege *sumu
degemoomoo
TURTLES
*kea
*fonu
SHELLFISH SPECIES
Species
Tvidacna
maxima
Hippopus
hippopus
Cassis
aornuta
nautilus
Lambis
lambis
Lambis S D .
Atrina
vexillum
Gafrarium
peotinatum
Asaphia
violascens
Codakia
sp.
Area ? sp.
Cardium
? sp.
Cardium
sp.
Spondylus
varians
Terebra
aveolata,
T.
arenulata
Name in text
su kima pahua
su kima kima
te pu loto waihanga
te (h)atili
te waeanga
te waeanga poto noua
te wolo
te pipi
te pipi mea
te pipi ta matau
te koeho
te pipi ta matau
te hala
te kai nap
te pu ka tolotolo
Lexicon
baehua
gima
buu
(not present)
waeanga
(not present)
wolo
bibi
bibi
bibi
goeho
bibi
(not present)
(not present)
buu
ADDITIONAL SHELLFISH FROM KAPINGAMARANGI LEXICON (Lieber and Dikepa 1974)
pearl oyster
baa
univalves in general
buu
Melampis
luteus
buu-hau
trumpet shell
buu-iliili
mollusc sp.
golo-golo
Vermetid gastropod
unga-goo
Nerita
pioea
unga-hihi
Nerita
polita
bulolo
Turbo
spp.
alili
PPN Term
r
puupuu
[
pipi
r
pipi
'pipi
f
pipi
c
puupuu
*paa
*puupuu
*puupuu
*puupuu
*'ungakoo
* 'unga-+*sisi
*palolo
*'alili
o
I
PLACE NAMES APPEARING IN THE TEXT AND IN
(1965) AND THEIR LEXICON EQUIVALENTS
Emory
Lexicon
Emory
Lexicon
kapingamarangi
torongahai
ringutoru
rikumanu
turuaimu
pepeio
nunakita
hukuniu
parakahi
werua
touhou
taringa
pungupungu
matiro
matuketuke
ramotu
sakenge
matawhei
gabingamalangi
dolongaahai
lingudolu
ligumanu
duluaimu
bebeioo
nunagida
huguniu
balagahi
weelua
touhou
daalinga
bungubungu
madiiloo
madugeduge
laamodu
tagenge
matawwei
hukuhenua
hepepa
tipae
tetau
nikuhatu
takairongo
tangawaka
hare
herengaua
herekoro
tirakau
tariha
tiahu
tokongo
tiraukame
pumatahati
matukerekere
huguhenua
[hebeba?]*
di bae
teedau
niguhadu
tagailongo
tangaawaga
hale
helengaua
halegoloo
di laagau
taliha
di ahu
togongo
di laagau mee
bumadahadi
madugelegele
*not in Lieber's list
-107-
APPENDIX 2
MOLLUSC SURVEY ON KAPINGAMARANGI
Graeme Ward
INTRODUCTION
In order to obtain some indication of the range of shellfish available to the prehistoric inhabitants of Kapingamarangi,
informal littoral surveys were conducted during the expedition.
Since Touhou and Werua are the main inhabited islets, the area
around them was the initial focus of the study. It soon became
clear, however, that neither the ocean nor lagoon shores of
either were areas rich in molluscs; consequently, attention
was shifted further afield. Some comments on present and past
exploitation are included.
METHODOLOGY
Since the aim was to obtain data concerning the range of
shellfish likely to have been available in the past, semi- or
non-quantitative survey techniques were used.
On the areas of the ocean reef east of Touhou and Werua,
two transects were walked at low tide between the shore and
the reef-edge. On each, populations of molluscs within a given
area were noted, and some collections made, at set distances
along each transect.
It was not possible to cover a similar distance offshore
from these or any islets on the lagoon side, since the water
deepened more quickly with distance from the shore. Nevertheless,
several areas up to 50 metres off the lagoon shore of Touhou
were investigated at low and intermediate tides using basic
diving gear.
Subsequently, the lagoon shores of several other islets
and inter-islet channels were inspected by walking the littoral
and diving off-shore to a distance of 100 metres.
Finally, much of the central, northern and southern
lagoon-side reef areas were systematically searched by crews of
up to seven divers on three occasions.
OCEAN SIDE REEF FLATS
Two transects of the reef flat oceanward from Touhou and
Werua were made on foot at low tide in the company of one older
and one younger Kapingi man. On each transect, at approximately
each 10 metres, note was made of the immediate environment and
• -108a collection, or at least identifications, made of all molluscs
in the vicinity of a 4 square metre area. The results of each
study are summarized below.
Transect 1, Northeast of Touhou Islet
Shore: Substrate of exposed coralline rock with large coral
rubble, steeply sloping from eroded islet edge. Some wall
building in progress, using coral heads and larger coral
rubble. No molluscs.
+ 10 metres: Smooth coral limestone substrate with scatters of
small pebbles of coral gravel. No molluscs observed.
+ 20 metres: Patch of coral gravel about 8 x 4 m exposed at low
tide, resting on coralline rock substrate. Under rubble from
100 to 300 mm in diameter were found three small gastropods,
Conus and Cypvaea spp. Nearby, a black, feathery-limbed
starfish and several crabs were found.
+ 30 metres: Water-sculptured coral limestone flat under 20 cm
water. Under flat coral stone c.20 cm in diameter were
found a small Cypvaea sp., a Mitra sp. and a sea urchin.
+ 40 metres: A slightly raised coral rock flat with algal growth
and scattered coral gravel. No molluscs observed.
+ 50 metres: A similar situation to that of the previous point.
At the base of a coral stone were two small cowries
{Cypvaea spp.). Also in the vicinity were several cone
shells occupied by hermit crabs.
+ 60 metres: Same as last two areas but with a slightly more
broken substrate. No molluscs observed.
+ 70 metres: As for previous areas but with a greater density of
coral gravel. This area produced several concentrations of
small molluscs from under coral rocks. A group of several
tiny gastropods, five Cypvaea annulus,
and several small
Conus sp. were found.In the same situations were hermit
crabs inhabiting a variety of gastropod shells, including
Mitva,
Stvombus
and Dvupa spp. This area was otherwise
notable for the presence of a number of small black and
white speckled sea cucumbers.
+ 80 metres: Same substrate. One loose coral stone but no
molluscs. Several Cevithium
nodulosum
shells inhabited by
hermit crabs.
+ 90 metres: Similar substrate. Two Cypvaea
annulus.
+ 100 metres: Similar substrate but lacking larger coral rocks.
No molluscs; some sea cucumbers.
+ 110 metres: Similar substrate but undulating; no large rocks,
and no molluscs.
+ 120 metres: As above.
-109+ 130 metres: Similar substrate. Nothing found under coral rocks.
Sea cucumbers in the survey area.
+ 140 metres: Similar area; large stone had algae at base.
Nothing else but large brown and large black and white
sea cucumbers.
+ 150 metres: A slightly higher area only 25 metres from the
reef edge with a scatter of fine gravel over the coral
limestone which is occasionally pitted with holes of up to
4 0 cm in diameter. In one of these, among the light gravel
filling it, were found two Cypraea
annulus.
+ 160 metres: Middle of raised area representing the algal crest
of the reef edge and containing isolated patches of algae
and bright corals being alternately covered and exposed by
shallow water; substrate broken and uneven. Small gastropods
including Cypraea spp. and Turbo sp. were found in holes.
+ 170 metres: Substrate of solid coral limestone, strongly
water-worn with clefts overwashed with foaming water; no
gravel or coral rocks. Only sea cucumbers seen.
Reef edge at commencement of downslope: More broken substrate
with a larger proportion of live coral. Several colourful
gastropods including Turbo and Cypraea spp. but mostly small
individuals. Larger Turbo argyrostomus
and Tridaana
maxima
were seen further out. Many small crabs at the reef edge.
The area approximately 20 metres back from the reef edge
appeared to be that exhibiting the greatest concentration of
molluscs in readily accessible situations on what was generally
a disappointingly barren reef flat. It was decided to obtain a
greater sample of this area.
+ 160 metres from shore: Raised area of reef flat, broken
substrate with gravel and coral rocks. At the base of a
coral head Drupa sp. and other small gastropods were seen.
+ 160 metres: Fifty metres further north on similar substrate.
One large Cypraea and a small Conus sp.
+ 160 metres: Forty metres further north the ridge was slightly
higher and less awash. Substrate pitted with holes.
Collected four Cypraea spp., small Tectus
and Turbo spp.
and a live Mitra sp. from a rubble-filled hollow.
+ 160 metres: Another 50 metres further north. Similar substrate.
Observed several small gastropods including Vasum,
Drupa
and Cypraea spp.
These last sample areas were situated opposite the shallow
channel between Touhou and Werua. The second transect commenced
toward the reef edge approximately 200 metres further on, near
the southern extremity of Werua Islet, and proceeded in towards
the shore. The reef flat was only about two-thirds of the width
of the flat at the place of the first transect.
-110Transect 2, Southeast of Werua Islet
+ 110 metres: At the reef edge in the surge zone; water 50 cm
deep. Substrate broken. Collected one large Turbo
argyvostomus.
+ 100 metres: Water-worn coral limestone substrate. Collected
large Conus sp. and small Tridaona
maxima.
+ 90 metres: Slightly higher area of ridge coloured by light
green algae. Several gastropods present, including small
Troahus maeulata,
Turbo,
Cypraea and Conus spp. Water too
deep for closer search.
+ 80 metres: Separately ridged area, pocked with hollows containing rubble. Hollows produce tiny gastropods and sea
cucumbers.
+ 70 metres: More level substrate, less hollows and rubble. Some
tiny gastropods.
+ 60 metres: Similar substrate, some larger coral rocks. Under
one rock were small Conus and Cypraea spp.
+ 50 metres: Similar substrate, some algae. Several tiny
gastropods.
+ 40 metres: Similar area. No molluscs.
+ 30 metres: Similar area. Two tiny gastropods under coral.
+ 20 metres: Even substrate of coral limestone, some scatters of
fine rubble. No molluscs.
+ 10 metres: At edge of raised coral limestone protruding into
shallow water. No molluscs, some crabs.
Shore: Comprising predominantly coralline sand pockets among
raised limestone. No bivalves or other molluscs found.
The reef flats on the ocean side of Touhou and Werua were
remarkably unproductive of edible shellfish. This is in keeping
with the paucity of molluscan habitats on the reef flat. The
inner shore and supra- and inter-tidal zones comprised largely
coarse gravels unsuitable for filter feeders and these had been
disturbed and, in some areas, extensively redeposited by the
recent storm waters. There were some coral rocks in this zone
at the south of Werua but these too lacked the expected range of
small gastropods such as Nerita
spp. and the various limpets
common to such areas elsewhere in the Pacific. On the reef flat
itself, only small deposits of fine gravels and sands were
located and the substrate of hard coralline rock was generally
exposed. There were few living corals or algae areas. There were
no beds of sea grass providing suitable habitats for bivalves and
small gastropods. Further out, in areas covered by water at all
but the lowest tides, little live coral was present; again, few
suitable molluscan habitats are present. Only at the small algal
crest and the donwslope of the reef edge were suitable areas and
these were relatively inextensive or difficult of access, even at
-Illlow tide. While turbulent conditions prevented a
inspection of the outer margin of the reef edge,
gastropods and Tvidaona
maxima were collected or
reef margin. It seems likely that this area is a
rich molluscan habitat but might not always have
for exploitation.
detailed
several large
observed at the
relatively
been available
No similar systematic inspections were made of the ocean
side reef flats of other islets of the Kapingamarangi Atoll. The
possibility remains relatively untested, therefore, that the
paucity of molluscs observed on the Touhou and Werua transects
was a function of over-exploitation or some other factor related
to these islets in particular. However, while a greater variety
and denser populations of gastropods and some bivalves were
observed on islets to the north and south, casual observations
suggested that most lacked extensive areas of suitable habitats
also. A notable exception were the areas of exposed coralline
reef rock found on Nunakita and several of the smaller southern
and northern islets; these areas had relatively dense populations
of small gastropods in their intertidal and suprasurge zones,
mainly Nerita
spp., small Turbo sp. and some limpets.
LAGOON SIDE LITTORALS
The lagoon littorals of Touhou and Werua were found to be
relatively poor in molluscan resources. In many places the sandy
shore was built on with coral rock foundations for houses. In
other places, as on the shore opposite canoe houses, the sand is
frequently disturbed. Along much of the shore, especially of
Touhou, the water is relatively polluted, the substrate silty
and overlain with debris. Nevertheless, there are areas of fine
sand and some seagrass among which some small bivalves were
observed.
The molluscan habitats of the lagoon shore were better
observed along some of the other islets. Along the southern shore
of Ringutoru, for example, there were extensive areas of fine
coralline sand and patchy sea grass as well as large clumps of
living coral. The shore beaches provided habitats for small
bivalves including Donax and Tellina
spp. Offshore, in sandy
bottom areas and in patches of sea grass were the bivalves
Asayhis,
Gafrarium,
Quidnipagus,
Codakia,
Soutaroopagia,
Spondylus
and Cardium spp. - all used throughout the Pacific as
food - and a variety of gastropods including Cerithium,
Cypraea,
Oliva,
Tevebra,
Mitra spp., as well as various murids and drupes.
A similar range of bivalves and gastropods was observed in the
shallow water offshore from Hare and several others of the
southern islets.
Generally, these habitats were easily accessible and some
of the species occurred in patches in considerable number,
especially the sand-dwelling filter feeders,Donax and
Tellina
spp. Elsewhere in Oceania, these small molluscs are collected in
large numbers and provide a dependable staple food resource.
They were not observed being consumed on Kapingamarangi, and
informants indicated that only the larger clams, Tvidaona
maxima
and Hippopus
hippopus_, and occasionally the larger Turbo
argyrostomus
were used as food today.
-112INTER-ISLET CHANNELS
The channel between Tetau and Nikuhatu islets was inspected
in some detail. It showed little sign of being a substantial
molluscan habitat except toward its lagoonward extremity, where
greater similarity to the inner reef areas were noted.
INNER REEF MARGINS
The larger part of the lagoon reef margin of Kapingamarangi
atoll comprises not sandy islet shores but intertidal reef.
These areas, extensive in the north, west and southwest of the
atoll, and complemented by some perpetually but shallowly
submerged reefs in the centre-west and southwest of the lagoon,
are extremely rich marine ecosystems. A considerable range of
molluscan habitats are found here, including those of the larger
molluscs. Among the significant gastropods found were large cones,
cowries, Trochus maculata
and Turbo spp., Lambis and
Strombus
spp. and the conspicuously burrowing Mitra and Terebva
spp.
Most significant of the bivalves were the large clams, Hippopus
hippopus
and Tvidaena
maxima. The latter were frequently found
partially embedded in coral interstices along the reef margins
to several metres in depth. Several areas of these reef margins
and the submerged inner reefs were surveyed.
CONTEMPORARY SHELLFISH EXPLOITATION
Shellfish does not appear to be a very significant food
resource to contemporary people on Kapingamarangi. Women and
children were not seen collecting small bivalves on the sandy
shores, or Nerita
spp. and other small gastropods from outcrops
of coral rock, or exploiting the variety of small molluscs
inhabiting the sandy and sea grass flats, as they so commonly
are on other tropical Pacific islands. No recent middens of
discarded shells were seen. It is unlikely that seasonality or
matters of access to suitable areas were involved in influencing
the apparent lack of use of this potentially important resource.
The only shellfish gathering which was observed was that
done by men. For example, an elderly informant collected Turbo
sp. shells from the reef edge and prepared their flesh for
consumption; and an acknowledged expert fisherman collected tiny
cowries from the reef flat for use in modern ornamental handicrafts. These examples, however, are insignificant when compared
with the economic importance of clam shells. In particular,
Tvidaena
maxima was a prime target for predation as a food source.
The large gastropod Turbo argyrostomus
was one of the few
molluscs claimed to be used for food today on Kapingamarangi
(cf. Buck, 1950: 49). To prove his claim that this species
collected at the reef edge was a delicious foodstuff, the elder
of the two informants assisting in the transects of the reef flat
proceeded to cook them on the shore. The shells were placed
upside down upon a coral rock slab resting on the ground. A fire
was kindled on top of them using dry coconut fronds. After it had
burned for three or four minutes, he used a palm midrib to remove
the shells from the fire. Each was hit against a nearby stone to
knock the operculum from its resting place without damaging the
shell mouth. The hot shell was held in one hand with the aid of
-113coconut fibre, and a small stick used in the other to prize out
the flesh. The small piece of gut was removed and only the muscle
was eaten. The flesh was succulent and tasty.
Collection of the two large clams, Tridacna
maxima and
Uippopus
hippopus
(cf. Buck, 1950: 49), was observed on several
occasions at the inner reef margins. Divers used a knife, quickly
and deftly thrust between the open valves of the clam to make it
release its hold on the substrate; another twist or two of the
knife within the shell would separate the mollusc from the two
valves; these would tumble away into deeper water as the diver
rose to the surface with the mollusc flesh. The gut was removed
from the adductor muscle and fleshy mantle with the teeth and
the flesh washed in sea water for later consumption.
These clams were a favourite, easily obtained and prepared
midday meal for fishing crews. Limes brought in the canoe were
squeezed into a bowl to marinate the mollusc flesh. Left for an
hour or so while the serious fishing proceeded, they provided an
easy luncheon for a fishing group who would not have to land on
shore or prepare a fire.
In view of the area of reef at Kapingamarangi atoll, it
appears unlikely that casual exploitation at this level would
ever produce a shortage of clam meat.
While it seems that Tridacna
maxima is the major, if not the
only species frequently sought for consumption as food today,
other shells, especially the small cowries, Monetavia
moneta and
Cypvaea annulus
are collected and used in ornamental handicrafts
of the type offered for sale in Ponape and elsewhere.
Buck (1950: 49) listed five molluscs as being of economic
importance on Kapingamarangi: pahua,
kima, avivi,
pu, waenga; he
also mentioned another: wovo.
These are Tridacna.
maxima,
Hippopus
hippopus_, Turbo argyrostomus,gastropod
generally, and small
members of the Strombidae family; and Atrina
vexillum.
Of these,
the last is rather unlikely as it is extremely rare. In addition,
Hippopus
hippopus
is barely considered even as poor food today.
This list of shellfish of economic importance in recent times is
short by comparison with other areas in the Pacific.
-114-
APPENDIX 3
THE ANALYSIS OF FISH BONE FROM KAPINGAMARANGI
Foss Leach and Janet Davidson
Comparative material for Pacific Island fishes is not
readily available to assist the identification of archaeological
finds. Over the past few years, specimens of about 200 species
have been caught and prepared to try to rectify this situation.
Selected cranial bones are mounted on boards according to anatomy,
and are now available for use at the Otago Archaeological
Laboratories. The method of processing archaeological assemblages
of fish bones has been described in detail elsewhere (Leach and
Davidson, 1977; Leach, 1979) and need only be summarised here.
(1) The fishbone is sorted into
Dentary
Articular
Quadrate
Maxilla
Premaxilla
Special bones (particularly diagnostic)
Residue
and rebagged.
(2) Dentaries (and each anatomical class in turn) are tipped
out, sorted, and rebagged by species using the comparative
collection. The identifications are written on the bag
labels and entered in forms for the site.
(3) Minimum numbers are calculated for each unit in the
assemblage (in this case spits), taking no account of
mismatched left and right pairs. A number of special bones,
such as spines of Biodon
sp., occur in large numbers on
each fish, and although they are counted, they only give a
minimum number of 1 for the assemblage unit, regardless of
total number.
Experience over the years identifying fish bone from many
archaeological sites has consistently indicated that quadrates
and articulars add very little to the overall minimum numbers
in cases where a reasonable assemblage is available for study.
Consequently, although these bones are still separated and
rebagged, they only need to be identified in cases of small
assemblages. The Kapingamarangi assemblages were quite sizeable,
and these bones were therefore not identified. In addition, time
prevented the identification of the maxillae at this stage. It is
doubtful whether these will ultimately make any difference to the
results presented here. The special bones were of considerable
importance. In particular, pharyngeal clusters of Scaridae,
-115Coridae, etc. quite commonly yield far higher minimum numbers
than those from the standard cranial anatomy. The Kapingamarangi
assemblages were no exception to this rule. In addition, the
characteristic terminal caudal vertebrae of Thunnidae and
Katsuwonidae were far more numerous than cranial bones. Spines
were the main portions indicating the presence of Balistidae and
Diodontidae.
For the Kapingamarangi fish bone, this general identification
procedure was very successful, and bones from only two or three
species could not be matched to genus level, at least, in the .
comparative collection. Identifications of course vary in level
of certainty from one family to another. For instance, members of
the family Scaridae are very difficult to identify to genus, let
alone to species. In other families, on the other hand, many
species and genera have quite diagnostic features, and more
precise identifications can therefore be made. The identifications
written on bags often refer to the closest species available in
the comparative collection. For instance, Holoaentrus
cf. ruber
should be understood to mean that this species is the most similar
amongst available material, and that the genus is fairly certain.
By contrast, cf. Holoaentrus
ruber implies that although this is
the most similar specimen, the bone may belong to some other
genus, for example Ostiahthys.
This method gives somewhat uneven
initial results through the taxonomic order, but it is possible
to pool these later into families, for example, for comparative
purposes. It is worth recording the more precise identifications
as one proceeds, as these can also be of specific use later. It
will be noted that specimens of Thunnidae and Katsuwonidae could
not be distinguished. This is because identifications largely
relied on vertebrae and tail elements, and although these are
distinctive of the two families, they are very difficult to
identify further.
Elasmobranchs have not yet been identified. This will
eventually require X-rays of the vertebrae. At this stage a
minimum number of 1 is entered for any spit, regardless of how
many vertebrae or teeth were present.
The Kapingamarangi fish bone was processed in this way, and
the basic results have been pooled to genus level and tabulated
below. The taxonomy throughout follows that of Munro (1967) and
the family numbers are consecutive from 1 = Orectolobidae.
No attempt has been made at this stage to work out size
frequency characteristics of the more numerous fish, but several
particularly odd features were noted during analysis. Firstly, a
few very large specimens of Caranx ?sp. were in the assemblages.
Secondly, more than half of the specimens of Lethrinus
?sp.
identified were enormous in comparison with anything in the
comparative collection. Representatives of family Belonidae
(Long toms), on the other hand, are very small specimens,
probably indicating that patches of sea grass (their main habitat
where they catch small fish) were never more extensive on
Kapingamarangi than they are today. Finally, almost all specimens
of families Thunnidae/Katsuwonidae in the assemblages are rather
small, and some are very small.
TABLE 12. BASIC RESULTS OF FISHBONE ANALYSIS, TIROKI
Levels
Spits
IV
1 2
3
Elasmobranchs
1 1
1
Muraenidae
cf.Gywnothorax
spp.
3
7
6
Holocentridae
cf. Holocentrus
cf. Ostiohthys
esp. ruber
esp. murdjan
2
l
5
l
12
l
l
l
3
4
Sphyraenidae
cf. Agrioposphyraena
esp. barracuda
Thunnidae, Katsuwonidae
cf. Neothunnus maoropterus
cf. Katsuwonus
pelamis
and
2
Carangidae
cf. Caranx e s p . ignobilis3
melampygus and
sexfasaiatus
cf. Elegatis
bipi-nnulatus
Epinephe1idae
cf. Epinephelus/Cephalopholis
cf. Pleotropoma
spp.
Lutjanidae
cf. Lutjanus
cf. Maoolor
spp.
spp.
Nemipteridae
cf. Monotaxis
Lethrinidae
cf. Lethrinus
Mullidae
cf. Parupeneus
granooulis
spp.
spp.
1 7
spp
25
2
24
6
5
5 13
5
23
9
6
1
Table 12 continued
Levels
Spits
IV
1 2
Coridae
cf. Eerrrigymnus
melapterus
cf. Cheilinus
spp.
Scaridae
Siganidae
cf. Sigarcus
1
19
35
spp.
Diodontidae
cf. Viodon
1 1
Balistidae
Total
5
6
7
III
8
9 10 11 12
II
13 14 15 16
I
17 18 19
1
5
1 1
19
3
Total
1 1 2 3 1 1 1 1
87
1
2
e s p . hystrix
4
1
2
Acanthuridae
cf. Naso e s p . bvevirostris
Tetrodontidae
cf. Arothron
3
4
3
1
2
5
8
1
I
spp.
4
3
1
1
4
7
4
1
82 124
108 14
1
1
1
7 7 9
1
14
6 5 4 5
1
4
1
1
1
1 1
10
M
22
'
392
TABLE 1 3 . BASIC RESULTS OF FISHBONE ANALYSIS, MURI-HALAU (KA2), MINIMUM NUMBERS
Levels
Layers
IV
Elasmobranchs
Muraenidae
c f . Gyrmothorax
1
2
1
1
3
4
1
spp.
2
3
5
1
II
7
6
1
1
4
2
Holocentridae
cf.
Holocentrus
spp.
Thunnidae/Katsuwonidae
Carangidae
c f . Cavanx s p p .
c f . Elegatis
biipinnulatus
Epinephelidae
c f . Epinephelns/Cephalopholis
c f . Pleatropoma
spp.
Lutjanidae
c f . Lutjanus
spp.
Nemipteridae
c f . Monotaxis
granooulis
Lethrinidae
c f . Lethrinus
spp.
1
1
1
1
1
3
1
1
4
1
1
2
5
2
5
1
1
1
2
3
1
1
1
1
1
3
3
1
2
1
Coridae
cf.
Cheilinus
spp.
Scaridae
1
1
1
3
1
3
4
2
1
1
1
1
1
1
Acanthuridae
c f . Naso s p p .
Diodontidae
c f . Diodon
spp.
1
Balistidae
1
1
4
1
1
1
1
0
8
19
26
20
7
13
Total
TABLE 14. BASIC RESULTS OF FISHBONE ANALYSIS FROM PUTAU (KA3), MINIMUM NUMBERS
Levels
Spits
3
III
4 5
1
1
4
4
IV
1
2
Elasmobranchs
Muraenidae
cf. Gymnotkorax
Belonidae
cf. Tylosurus
spp.
esp.
1
6
1
7
1
3
2
8
II
9 10 11 12 13 14 15 16
1
1
5
3
1
1
1
1
1 1 2
1
1
1
indicus
1
Holocentridae
cf. Ostiahthys
cf. Holoaentrus
esp. murdjan
esp. ruber
Thunnidae/Katsuwonidae
cf. Neothunnus
macropterus
cf. Katsuwonus
petamis
1
1
Nemipter id ae
cf. Monotaxis
esp.
sebae
4
1
1
1
39
1
5
1
2
14
5
2
3
18
1
1
23
2
spp.
2
2
1
2
1
2
1
3
1
1
2
1
3
3
1
1
1
2
1
vivescens
2
1
1
1
1
1
2
1
1
1
1
1
2
1
1
1
1
19
8
2
5
5
3
1
2
2
1
3
1
6
1
1
1
1
3
4
39
18
2
1
1
1
1
1
1
18
3
2
2
2
17
2
1
1
coaerulaureus
granoeulis
H
I
Carangidae
cf. Caranx spp.
c f . Caranx e s p .
ignobilis
c f . Elegatis
bipinnulatus
Lutjanidae
c f . Lutjanus
esp.
c f . Macolov s p p .
c f . Aprion
esp.
1 1
1 1
and
Epinephelidae
c f . Epinephalus/Cephalopholis
spp.
c f . Plectropoma
1 2
1
Total
4
1
Exocoetidae
Caesiodidae
c f . Caes'io
1
1
1
I
17 18 19 20 21
1
1
2
2
2
1 1 1
Table 14 continued
Levels
Spits
Lethrinidae
c f. Le thrinus
IV
1 2
3
4
III
5 6
7
spp.
8
II
9 10 11 12 13 14 15 16
4
1
2
I
17 18 19 20 21
Total
14
3
Mullidae
cf.
Parupeneus
Coridae
c f . CheiVinus
spp.
spp.
Scaridae
3
1 4
1
1
2
1 1 1 2
1 1
11
1 1 1
20
3
4
3
5
3
8
1
8
4
1 1 1 2
Acanthuridae
cf. Naso spp.
76
3
Scorpaenidae
cf. Soovpaena
cardinal-is
Diodontidae
c f. Diodon spp.
Tetrodontidae
cf. Arothron
Balistidae
Total
1
to
o
1
1
1
1
1
1
1
1
1
10
1
spp.
6
6 37
1 2
2
1
24 15
6 11 12
6 2
1
1
2
1
4
42
11
59 20 12 15 26 34 20 16 48
15
3
1
3 6
389
I
TABLE 15: BASIC RESULTS OF FISHBONE ANALYSIS FROM NGEIHO-HEREU (KA4), MINIMUM NUMBERS
Levels
Spits
Elasmobranchs
IV
1 2
3
III
4 5
1 1
1 1 1 1
6
7
II
9 10 11 12 13 14 15 16 17 18 19
8
1
1
1
1
1
1
1
1
1
Total
1
16
Muraenidae
cf. Gymnothorax
Angui11i forme s
spp.
2 3
1 1 3 4
? Family, species A
? Family, species B
? Family, species C
4 1 4 2 4 1 2
1
2
33
11
2
1
11
1
2
1
Belonidae
cf. Tylosurus
Hoi oc en t r id ae
cf. Holocentrus
cf. Ostiahthys
indious
1
e s p . ruber
spp.
Acanthocybiidae
cf. Aoanthoaybium
4
2
3
2
1
1
2
1
1
1
18
1
1
M
solandri
Thunnidae/Katsuwonidae
cf. Neothunnus macropterus
cf. Katsuwonus
pelamis
1
and
1 2
Carangidae
cf. Caranx melampygus
cf. Caranx e s p . ignobilis
cf. Carangoides
oblongus
cf Carangoides
latiaaudis
1 2 1 1
2 1 1
2
1
1
1
1
1
2
1 3 1 2
21
1
1
3
1
4
9
1
1
1
1
1
Epinephelidae
cf. Epinephelus/Cephalopholis
cf. Pleotropoma
spp.
L u it j a n i d a e
cf. Lutjanus
cf. Aprion
i
e s p . sebae
se
virescens
spp.
2 33
1 22
22
1
1 22 1 1 4 4
1 22
3
11
1
33 11 22 3 3 3 3 4 4 3 3 2 22 2
1 22 11 11 1 1 1 1 1 1 1 1
3
34 4
2 2 2 2 33
1
11
1
1
39
15
19
3
'
Table 15 continued
Levels
Spits
IV
1 2
3
4
III
5 6
II
9 10 11 12 13 14 15 16 17 18 19
7
8
1 1 3
Total
Nemipteridae
cf. Monotaxis
granoaulis
l l
1
2
3
spp.
2 1
1
1
1 1 1 1
1 2
19
1 •2
Lethrinidae
cf. Lethrinus
10
Mullidae
cf. Mulloidichthys
auriflamma
l
Kyphosidae
cf . typhosus spp.
1
Coridae
cf. Hemigymnus
cf. Cheilinus
spp.
Scaridae
2
1
melapterus
2 1
4
1
1
6
1
3
2
4 13
5
2
1
2
12
1
7
1
2
2
3
1
1
54
Siganidae
1
cf. Siganus
spp.
4
Acanthuridae
1
Diodontidae
cf. Diodon spp.
Tetrodontidae
cf. Arothron
spp.
1
1 1 1 1
1
2
Balistidae
cf.Rhinecanthus
aeuleatus
cf .Meliehthys
buniva
cf.Pseudobalistes flavimarginatus
ToiaZ
i 3
26 31
2
11
1
9 18 38
2
1 1
1 2
1
3
15
1
25 17 15 29 29 22 21 20 10
9
330
M
M
I
-123-
APPENDIX 4
KAPINGAMARANGI CRUSTACEAN REMAINS
Graeme Mason
The Kapingamarangi crustacean remains all belong to the
Order Decapoda Suborder Reptantia, the super sections Macrura
(crayfish), Anomura (hermit crabs etc.), and Brachyura (true
crabs) all being represented. The most frequently occurring parts
comprise cheliped and pereopod (pincher and leg) fragments among
which the dactylus is commonest.
By anatomy the remains are distributed thus:
Cheliped
Left dactylus
Right dactylus
Left propodus
Right propodus
5
11
0
3
Pereopod
All dactyli
Miscellaneous
11
4
Mandible
Fragments
2
Miscellaneous fragments
11
Total 47
The absence of comparative material and, at present, a number
of taxonomically important works, coupled with the non-existence
of modern reviews of many major groups (particularly the
Xanthidae, the largest single crab family in the Indo-Pacific and
one which contains many edible species) has severely limited
identification to a level well below what the writer would wish,
so this report must be regarded as preliminary only. It should
also be remarked that chelipeds and pereopods are rarely used as
taxobases.
Super Section Macrura
Family Palinuridae (spiny crayfish):
Super Section Anomura
Family Coenobitidae
Birgus
latro
(coconut crab):
2 mandibles
tip of propodus,right
cheliped
-124Coenobita
sp. (hermit crab): L cheliped dactylus
tip R cheliped propodus
Coenobitid:
3 dactyli
Super Section Brachyura
Family Xanthidae (Black-fingered crabs)
Xanthid spp.: 7 cheliped dactyli
The balance of the material may include some Portunid
crabs (swimming crabs) and a few small species whose identity is
at present in doubt.
The absence of the large edible terrestrial Gecarcinids
(especially Cardisoma
spp.) and Grapsids (Sesarma
spp.Geograpsus
spp.) is noteworthy as they are known at present on the island.
The scarcity of Birgus
remains is disappointing in view of the
light they may cast on human movement in the Pacific area. In
this context, Cardisoma
longipes
(Gecarcinidae) has a rather
patchy Pacific distribution which may be perhaps partly
explained (it is large and palatable) by human transport and
maybe, over exploitation locally.
Use of terrestrial, littoral and sublittoral crustaceans
is clearly demonstrated. The distribution of crustacean remains
is shown in the table below.
Table 16.
Crayfish
Coconut crab
Hermit crab
Xanthids
Crustacean
Crustacean remains from Kapingamarangi
IV
III
II
I
K M KA2 KA3 KA4 KAl KA2 KA3 KA4 KAl KA3 KA4 KAl KA3
2
1
2
1
1
7
2
1 4
2
3
1
5
9
4
2
-125-
APPENDIX 5
ANALYSIS OF ARCHAEOLOGICAL SHELL
FROM KAPINGAMARANGI
Simon Holdaway
The modern comparative material collected by Leach and
Ward during their expedition was sorted into taxonomic groups
and identified with the assistance of Cernohorsky of the
Auckland Museum. This material, together with the existing
comparative collection of tropical shells already in the Otago
Archaeological Laboratories, formed the basis for identifying
the archaeological residues from Kapingamarangi.
Once the shells had been separately bagged, each bag was
sorted into pieces relevant to minimum number estimates, and
residue. The latter was rebagged for storage. Diagnostic
material was sorted into eight categories and rebagged.
Donax spp.
Lambis
spp.
Tvidaona
spp.
Turbo
spp.
Strombus/Mitra/C'onus/Trochus/Drupa
spp .
Other Gastropods
Other Bivalves
Cypraea
spp.
Each of these categories was later separately treated to
obtain accurate identifications, and minimum numbers.
Some comments on the main individual shells found are
appropriate, since their treatment varied somewhat.
Turbo argyrostomus:
Here, both the number of opercula and
the number of shell apexes were counted and recorded. The
minimum number per square and spit was the larger of the two
numbers. Operculae were counted when half or greater was present.
It is assumed that all individuals are Turbo
argyrostomus,
although it is possible that some of the operculae are from
other Turbo species.
Rhinoolavis
asper:
Minimum numbers were calculated from the
apex. Again, whole shells were more easily identified.
Strombus
luhuanus:
The apex was used for minimum number
calculations; all other fragments were ignored.
Lambis spp.: Only the apex was used for minimum numbers.
Separate fragments containing siphonal appendages which allow
-126identification, but not including an apex were not included for
minimum number calculation. It is noteworthy that there are very
few complete shells of Lambis spp.
Cypvaea spp.: All shells were identified to species level
where possible. Only complete or near complete shells were used
for minimum number calculation. Most specimens were very small,
and the majority were probably not collected as food.
Drupa spp.: The apex was not always very clear, so only
complete or near complete shells were used for minimum numbers.
Mitra
spp.: The apex was used for minimum numbers.
Conus spp.: The apex was used for minimum number calculation.
Many of the specimens, especially the smaller ones, are unlikely
to have been of economic importance.All, however, were identified
where possible. As species identification relies on the aperture
and siphonal canal shape, there is a bias toward complete shells
in the minimum number calculations. Individuals identified only
to the level of Conus sp. were not used for minimum numbers, as
these were invariably worn, consistent with a beach sand origin.
Terebra
spp.: The apex was used for minimum numbers. The
whorl shape is quite diagnostic, so shell fragments could also
be identified.
Other Gastropods: The majority of shells in this category
were Nerita
spp. Whole shells or nearly whole shells were used
for minimum numbers. Most of the Nerita
were identified to genus
level only, as there are many species of Nerita
for which
comparative material was not available. Minimum numbers of other
gastropods were assessed from either the apex or whole shells.
Atrina
sp. and Pinotada
sp.: Very few fragments of these
were found, but because of their importance were booked as
present in any spit in which even tiny pieces were recovered.
This includes the few artefactual pieces.
Tridaona
spp. and Hippopus
sp.: Identification to genus
level was achieved for nearly all fragments. Where complete
valves were present, species identification and valve siding were
possible. Minimum number estimates are based on valves with a
hinge present.
Donax spp.: Identification was to genus level only. Minimum
numbers were calculated as half the number of shells present
without attention to left or right valves. Broken valves were
counted only when the hinge was present.
Other Bivalves: All bivalves were identified to species
where possible. Minimum numbers were calculated as half the
total number of valves with hinges (or near complete valves). In
the case of broken valves, an individual was counted when the
majority of the hinge was present. It should be noted that dead
specimens of Glyoydonta
mavioa are common in coral sand and not
all those in the sites are considered to be economic in origin.
-127The full set of tables of these identifications is a
sizeable document, and for convenience only a summary is
presented here. In Table 17, the actual species identified are
listed in their appropriate taxonomic position. In Tables 18-21
the archaeological shells are grouped according to family for
ease of comparison and economic assessment. In view of this
summary treatment here, some additional comments should be made,
It is notable that no specimen of. Troohus
nilotious
was
recovered on Kapingamarangi, either archaeologically, or in the
modern collection. Troohus
sp. generally appeared to be rare.
Figures for Strombidae, of course, include both Strombus
spp.
and Lambis sp. Of these, the former were four times as numerous,
In the family Tridacnidae, only two species are positively
identified - Tridaona
maxima and Hippopus
hippopus.
Apart from
two small fragments of Spondylus
varians3
all material in this
family was Spondylus
duoalis.
TABLE 17: THE SHELLFISH SPECIES IDENTIFIED FROM KAPINGAMARANGI
ARRANGED IN TAXONOMIC ORDER FOLLOWING CERNOHORSKY
(1972: 25-30)
Modern
27 Trochidae
Troohus
Troohus
Troohus
maoulatus
histrio
*
*
sp.
2 9 Turbinidae
Turbo argyrostomus
34 Neritidae
Nerita
polita
Nerita
undata
Nerita
sp.
81 C e r i t h i i d a e
Cerithium
aluoo
Cerithium
oolumna
Cerithium
nodulosum
Rhinoolavis
asper
Rhinoolavis
sp.
90 Strombidae
Strombus
luhuanus
Strombus
mutabilis
Strombus
gibberulus
gibbosus"
Lambis chiragra
ohiragra
102 Cypraeidae
Cypraea
depressa
Cypraea
argus
Cypraea oarneola
Cypraea
annulus
Cypraea•
caputserpentis
Cypraea
sourra
Cypraea
isabella
Cypraea
poraria
Cypraea
teres
Cypraea
lynx
Cypraea
testudinaria
Cypraea arabioa
(juvenile)
Cypraea mauritiana
(juvenile)
Archaeological
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
-128Table 17 continued
Cypraea
sp.
Monetaria
moneta
Monetaria
sp.
108 Naticidae
Poliniaes
melanostomus
Poliniaes
sp.
109 Cassidae
Casmavia
erinaoeus
Malea pomum
110 Tonnidae
Tonna perdix
112 Cymatiidae
Cymatium
nicobaricum
Cymatium
sp.
Distorsio
anus
,
113 Bursidae
Bursa
granularis
Bursa
sp.
116 Muricidae
Morula
granulata
Chiooreus
brunneus
Drupa morum
Thais
armigera
Thais
tuberosa
Thais
sp.
117 Magilidae
Quoyula
madreporarum
119 Buccinidae
Cantharus
undosus
123 Olividae
Oliva
miniaaea
124 Vasidae
Vasum sp.
126 Harpidae
Harpa amouretta
129 Mitridae
Mitra
stiotioa
Mitra
mitra
Mitra
litterata
123 Conidae
Conus lividus
Conus floridulus
Conus arenatus
Conus ebvaeus
Conus ranunculus
Conus tulipa
Conus sponsalis
Conus pulicarius
Conus miles
Conus flavidus
Terebridae
Terebra
affinis
Terebra
areolata
Terebra
sp.
Terebra
crenulata
Terebra
maoulata
modern
*
archaeological
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
-129Table 17 continued
135 Architectonicidae
Eeliacus
variegatus
239 Ellobiidae
Melampus
flavus
315 Arcidae
Barbatia
deoussata
321 Pinnidae
Atrina
(Atrina)
vexillum
322 Pteriidae
Pinctada
margaritifera
Pinctada
maculata
(juvenile)
Pteria
penguin
Pteriidae ? genus
326 Pectinidae
G'Ioripallium
pallium
33 0 Spondylidae
Spondylus
ducalis
Spondylus
varians
354 Carditidae
Beguina
semiorbiculata
357 C a r d i i d a e
Cardium
elongatum
358 Tridacnidae
Tridacna
maxima
Tridacna
sp.
Eippopus
hippopus
360 Mesodesmatidae
Atactodea
striata
365 Donacidae
Donax sp.
366 Tellinidae
Quidnipagus
palatum
Scutarcopagia
scobinata
367 Psammobiidae
Asaphis
violascens
370 Semelidae
Semele
sp.
372 Trapeziidae
Trapezium
bicarinatum
380 Veneridae
Gafrarium
pectinatum
Glycydonta
marica
Periglypta
reticulata
Periglypta
puerpera
modern
archaeological
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
TABLE 1 8 . MINIMUM NUMBERS OF SHELLFISH FROM TIROKI
Levels
Spits
Trochidae
Turbinidae
Neritidae
Cerithiidae
Strombidae
Cypraeidae
Cassidae
Cymatiidae
Muricidae
Magilidae
Buccinidae
Mitridae
Conidae
Arcidae
Pinnidae
Pteriidae
Spondylidae
Carditidae
Cardiidae
Tridacnidae
Mesodesmatidae
Donacidae
Tellinidae
Psairanobiidae
Total
IV
1
2
I
3
8
1
14
2
22
3
9
2
4
4
1
4
4
1
1
1
1
(KA-1)
III
8
5
2
15
1
31
52
5
2
I
9
10
11
12
46
1
16
57
65 301
1
1
1
3
6
1
15
4
13
II
14
91
3
6
16
56
1
22
12 181
1
3
1
1
1
1
1
2
4
1
1
6
1
1
1
1
1
1
1
2
4
5
25
1
2
3
5
4
4
1
8
1
20
42
62
1
3
18
1
8
13
21
38
68
54
20
2
62
76 378
4
1
12
2
10
2
Tot,
1
2
1
1
1
13
2
8
1
1
34
1
1
1
24 241
22
703
1
1
2
1
1
1
4
2
2
2
1
2
1
12
1
I
17
1
516
15
1
19
20
1
1
1
3
1
3
4
19
2
3
1
1
6
12
1
69
1
2
1
1
E
15
119
4
62
9
36
1
1
3
1
2
o
I
TABLE 19. MINIMUM NUMBERS OF SHELLFISH FROM MURIHALAU (KA-2)
Levels
Layers
Trochidae
Turbinidae
Neritidae
Cerithiidae
Strombidae
Cypraeidae
Naticidae
Cymatiidae
Muricidae
Magilidae
Mitridae
Conidae
Terebridae
Ellobidae
Arcidae
Pteriidae
Spondylidae
Cardiidae
Tridacnidae
Donacidae
Psammobiidae
Trapeziidae
Veneridae
Total
1
2
45
2
1
27
7
76
59
1
1
5
1
2
1
3
41
17
1
4
66
1
1
4
1
5
1
IV
4
2
219
17
3
46
1
5
6
36
1
33
5
6
9
8
6
2
3
5
2
1
1
2
401
49
2
21
262
1
1
4
4
6
17
5
III
7
1
52
5
1
3
453
54
2
37
275
1
2
4
4
6
17
5
1
5
1
9
5
11
506
2
1
9
154
1413
16
13
1
i
1
2
2
1
4
1
2
87
1
41
2
63
1
5
2
1
227
143
210
1
2
1
176
2
780
2
40
102
1
43
5
1
6
4
7
450
2
1
8
97 1259
total
3
1
4
56
U)
h-1
I
TABLE 20. MINIMUM NUMBERS OF SHELLFISH FROM PUTAU (KA-3)
Levels
Spits
iv
1 2 I
Trochidae
30 26 56
Turbinidae
2 3 5
Neritidae
Cerithiidae
1 1
Strombidae
1
1
Cypraeidae
Naticidae
Cymatiidae
Bursidae
Muricidae
Magilidae
Mitridae
Conidae
Terebridae
2 5 7
Arcidae
Pectinidae
Spondylidae
2 1 3
Cardiidae
1 1 2
Tridacnidae
Mesodesmatidae
Donacidae
Tellinidae
Psammobiidae
Veneridae
Total
ill
3 4 5 6
7
£
8
II
9 10 11 12 13 14 15 16 I
2
3
51
142
212
17
10 26 35 139 158
2 2 2
2
8
1
13
3
3
1
2 42
1 2
53 100
5 12
1
1
4
4
2
1
1
5
1
1
4
3
1
1
1
1
4
1
6
7
6
1
2
3
2
1
4
2
2
2
71
26
28
4
3
2
1
5
2
2
2
1
1
1
7
3
10
26
2
7
1
1
1
2
4
3
2
38
1
1
17
4
1
6
1
17
1
1
2
1
6
2
68
13
2
2
4
1
2
64
7
1
4
1
1
1
1
1
3
38 37 75 35 63 19 75 101 293 188 165 283 172 117
1
1
1
1
1
1
10
1
1
10
14 147 860
1
5 40
3
1
2
19
4 38
4
1
2
1
2
1.. 2
3
1
14
3
1
14
1
1
1
7
1
6
1
13
1
10 15 199
1
1
1
1 12
1
5
I
17 18 19 20 21 Z
1
66
1
1
1
37
2
11
1
6
1
1
1
1
3
2
3
9
13
1
1
5
1
1
4
1
10
1
57 222
4
3
16
1
3
90 472
3 13
1
1
1 32
1
2
1
7
1
6
1
1
1
13
1
1
1
59 65 33 178 1260 97 68 93 250
1
6
3
55
1
3
1
98 606
total
13
1527
66
4
121
83
5
4
2
3
2
4
20
3
38
1
14
15
23
1
261
2
16
6
2234
I
TABLE 21. MINIMUM NUMBERS OF SHELLFISH FROM NGEIHO-HEREU (KA-4)
LeveIs
Spits
Trochidae
Turbinidae
Neritidae
Cerithiidae
Strombidae
Cypraeidae
Naticidae
Cymatiidae
Magilidae
Vasidae
Harpidae
Conidae
Terebridae
Arcidae
Pinnidae
Pteriidae
Spondylidae
Carditidae
Cardiidae
Tridacnidae
Donacidae
Tellinidae
Psammobiidae
Veneridae
Total
1
27
IV
2
I
3
68
41
4
3
2
III
5
II
6
I
73 483 559
1
1
2
1
3
6
3
9
1
8
9
10
11
1
203 141 263 263 121
1
3
11
1
1
1
2
1 56 19
3
3
5
6 10
1
12
13
14
15
16
57
6
35
5
39
2
21
3
4
1
8
12
1
7
3
2
2
1
1
4
1
5
1
1
1
2
3
1
1
4
4
2
1
1
1
2
1
1
2
2
4
6
3
9
1
1
1
2
26
10
1
16
4
9
1
1
1
1
1
221 155 354 321 154
96
71
1
7
1
7
1
1
3
12
2
28
47
75
5
8
81 505 599
1
5
1
1
.
1
4
11:
32
3
79
59
1
2
1
1
1
10
1
2
1
3
4
E
1
1
1
17
1
1
1
1
3
4
1
1
1
1
1
54
34
10
3
4
1
2
14
95
1
5
4
1 1471
Total
1
1774
34
3
82
68
1
2
1
1
2
10
1
13
0
4
4
1
8
21
104
1
5
4
2145
I
h-1
U)
I
-134-
APPENDIX 6
SAND TEMPER IN A SHERD FROM KAPINGAMARANGI
W.R. Dickinson
A thin section of a sherd from Kapingamarangi atoll in the
Caroline Islands was sent for study by B.F.Leach. The temper in
the sherd is a volcanic sand of complex composition perhaps
derived from somewhere in the Solomon Islands, but the sherd is
possibly from ware introduced during this century from Japan.
The composition of the sand temper is approximately as follows:
15% quartz grains, 40% feldspar grains, 20% volcanic rock
fragments, 15% chert and microgranular rock fragments, 5%
pyroxene grains and 5% opaque iron oxide grains. This sand is
unlike any observed from Palau, Yap, or Truk. The presence of
quartz implies exposure of granitic or metamorphic rocks, or
sediments derived from such sources, in the provenance.
Petrographic Report WRD-8 3
-135-
APPENDIX 7
SAND TEMPER PETROGRAPHY OF A SHERD AND SOME JAPANESE TILES
FROM KAPINGAMARANGI
C.A.Landis
A solitary sherd collected from Kapingamarangi by B.F.Leach
was examined by W.R.Dickinson. In his report (Appendix 6 ) ,
Dickinson described the sherd temper as being of complex volcanic
origin, possibly derived from the Solomon Islands. In addition he
pointed out that the presence of quartz in the sherd (estimated
15%) "implies exposure of granitic or metamorphic rock, or
sediments derived from such sources, in the provenance". In view
of its exotic temper petrography, Dickinson also suggested the
possibility that the sherd could be "ware introduced during this
century from Japan".
Japanese tiles do occur on Kapingamarangi. In order to test
the possibility that the sherd may be such recent material, four
Japanese tiles from the island were studied petrographically and
compared with the Kapingamarangi sherd.
The four tile samples (79.425 to 79.428 inclusive) are
nearly identical. Their temper consists almost entirely of quartz
and alkali feldspar (estimated proportions 60:40 to 40:60).
The quartz occurs as very angular single crystals up to 1 mm in
maximum dimension, cracked but unstrained, and generally coarser
than the associated feldspar. Much of the feldspar is albitic in
composition and is marginally vitrified. The more highly
vitrified grains show no twinning and may be a potassic feldspar.
Minor amounts of bleached biotite and rhyolitic lithic fragments
are also present. In addition, two zircon grains were identified.
The temper of the Kapingamarangi sherd (79.1) is essentially
as described by Dickinson. In addition I would point out that
volcanic lithic grains (estimated 25% of total temper) are
largely siliceous in nature (rhyolite-dacite) and that some of
the associated quartz and feldspar is surely of the same
provenance. Other mineral grains (calcium-plagioclase and augite)
are also common and imply a more basic igneous provenance. A
source in the Solomons or New Guinea seems entirely plausible.
It is very likely that the sample originated in a Pacific margin
environment.
-136-
APPENDIX 8
PHYSICAL ANALYSIS OF KAPINGAMARANGI POTTERY
Michiko Intoh and Foss Leach
The single piece of pottery from Kapingamarangi is a rim
sherd from a large, near vertical-sided bowl with a plain rim.
The sherd varies in thickness from 10 to 12 mm. From the sherd
curvature, the internal diameter of the original rim is estimated
as about 16 cra. The paste is 'fairly fine textured, and the sherd
shows flashes of red and black, characteristic of a firing
atmosphere which is only partly controlled, such as occurs in
an open fire where both oxidising and reducing conditions prevail,
This feature gives the sherd the appearance of a typical specimen
of Oceanic pottery.
The sherd is exceptionally hard for an example of Oceanic
pottery. In this respect it is more similar to modern field tiles
or other European pottery vessels, and this origin must be
considered. Two obvious possibilities are early Spanish
pottery or a piece of Japanese waste-pipe or field tile. Sherds
from such pipes are quite common in Micronesia, and were exported
from Japan during the period of Japanese administration (see
Figure 42). Spanish pottery occurs in Pacific island archaeological sites, but its characteristics are not well known.
To shed further light on this, water absorption tests were
carried out on the sherd and other specimens for comparison.
Total volume of sand temper (%) was estimated by point counting
thin sections. All work was rigorously cross-checked against
known standards to achieve maximum reliability. The results are
given in the table below.
Table 22. Physical Analysis of Pottery.
Specimen
Dokan 79.425
Dokan 79.426
Dokan 79.427
Dokan 79.428
WA
11.13
10.93
12.16
11.88
11.53
±0.29
AP
22.27
22.02
24.09
23.58
22.99
±0.50
BD
2.00
2.01
1.98
1.98
1.99
±0.01
SG
2.57
2.58
2.61
2.60
2.59
±0.01
ST
26.8
25.2
32.0
25.0
27.25
±1.63
WT
11
11
11
11
Kahula Spanish
18.95
34.01
1.79
2.72
18.8
12
Kapingamarangi 79.1
10>5Q
Mean
and
SE of
mean
21>45
2>Q4
2>6Q
32<6
1Q_12
WA = water absorption (%), AP = apparent porosity (%), BD = bulk
density (gm/cc), SG = specific gravity, ST = sand tempering
(% v/v), WT = wall thickness (mm)
-137Is the sherd from a Japanese field tile ? The water
absorption tests show the specimen to have rather similar
characteristics to pieces of dokan waste-pipes collected on
several islets on Kapingamarangi. The degree of sand tempering is
also similar. However, the general appearance of these tiles is
completely different. Firstly, dokan are glazed on both inner and
outer surfaces; secondly, the curvature is quite different, with
an internal diameter of 7 cm; thirdly, the firing conditions of
dokan are highly controlled and give a uniform red colour. Origin
from a dokan tile can be ruled out with certainty.
Is the sherd from a piece of Spanish pottery ? Unfortunately
very little is known of the seventeenth and eighteenth century
Spanish pottery in the Pacific. Of material examined at first
hand, the glazed potteries can be ruled out, leaving a uniform
red, fine paste pottery,similar to the well known 'flower-pot'
type of modern European pottery. A sherd of this pottery from
Kahula in Taumako is included in the Table above for comparison.
It is clearly quite a different ware, being much more porous
than the Kapingamarangi sherd. In addition, Spanish pottery was
kiln fired, giving a well controlled firing atmosphere and
surface appearance. Spanish origin can also be ruled out.
Is the sherd a specimen of Oceanic pottery ? The water
absorption results for the sherd were not especially typical
of known Oceanic potteries. The hardness of the sherd is reflected
in the low apparent porosity of 21.45%. Pacific island potteries
range from mean porosity values of 28% to nearly 50% for
different areas, with Micronesian pottery generally being in the
upper half of this range. The closest pottery to the Kapingamarangi sherd is that from Yanuca in Fiji, rare specimens of
which can be as low as 20.16%. The values for other
characteristics in the Table are well within the ranges for many
Pacific potteries.
In conclusion, based on these simple tests the
Kapingamarangi sherd shows itself to be different from known
Japanese or Spanish pottery which occurs in the Pacific. It is
from a hard and well fired pottery vessel which has been fired
in a relatively uncontrolled atmosphere. It is unlike the biscuit
wares of Micronesia, and is more like an unusually hard specimen
of pottery from further to the south.
-138-
APPENDIX 9
KAPINGAMARANGI ROCKS
Graeme Mason
Being a coral atoll of some geological maturity,
Kapingamarangi is unlikely to have washed onto its shores
locally derived stone material, with the exception of young,
well cemented limestones. The source of non-bioclastic hard
rocks is assumed to be either material washed ashore in drift
wood or places beyond the island tapped by human agency.
Differentiation of material according to these alternatives
has not been attempted as only in one instance (79.551) is
there any evidence of human modification of a natural clast.
In shape, the clasts (ranging in size from 2 - 1 5 cm, most
commonly about 5 - 6 cm in greatest dimension) are generally
well rounded, but include a fair number which are sub-angular,
possibly as a result of heat fracturing and subsequent weathering.
In most cases the rocks have been identified only on hand
specimen characteristics, backed by cursory examination of some
thin sections. No detailed petrography has been attempted.
LIMESTONES
79.547. Ka-2, layer 2.
Burnt pebble of dense compact carbonate, grey-white
aphanitic, homogeneous; shows worm borings.
79.548. Ka-2, layer 2.
Yellowish white, dense, finely granular carbonate pebble,
showing irregularly nodulose surface. Under the hand lens, dark
sub-angular opaque particles to 0.25 mm set in grey-white
aphanitic groundmass. Microscopically, the texture suggests that
the rock may be a concretionary nodule developed in the lower
part of a soil profile.
79.552. Ka-2, layer 1.
Pale buff-white compact carbonate rock, finely granular,
containing black opaque grains to 0.25 mm. Sedimentary texture.
79.554. Ka-3, spit 20.
Mineralised coral fragment showing secondary calcite and
solution pitting.
79.556. Ka-3, spit 20.
Grey-white dense aphanitic limestone.
-13979.557. Ka-3, spit 20.
Grey-white dense aphanitic limestone
METAMORPHIC ROCKS
79.64. Part of large boulder washed ashore with log in 1978.
White granular compact rock showing secondary veins
containing minute well developed quartz crystals, minor amounts
of dark opaque grains. In thin section, well rounded quartz
grains are visible showing dusty margins and clear overgrowths
in optical continuity with the grains. Strain extinction
present in many grains, some boundaries sutured. Biotite
present in minor amounts.
79.83. Surface find, Sakenge Islet.
Fine grained, light grey-black mottled rock with fine
white veins prominent, sulphide particles (0.25 mm) visible.
Mylonitic structure present. In thin section, the specimen
is a slightly crushed, poorly sorted greywacke with clasts
up to medium sand size. Feldspars clouded with ragged plumes
of secondary minerals developed at their edges. The rock
appears to be a mylonized greywacke.
79.296. Part of modern pandanus anvil, Pepeio Islet.
Burnt,light grey,compact,coarsely granular rock, weathering
to rusty brown to chocolate brown, predominately sub-angular
to sub-rounded quartz grains up to about 2 mm, numerous minute
black opaque grains visible. Linear concentrations of small
dark minerals present. In thin section, shows granoblastic
texture, quartz grains comprising bulk of the rock. The
boundaries of the sub-eguant grains are sutured and contain
indeterminate dark brown opaque material, occasional irregular
patches of chlorite.The rock is a metamorphic quartzite.
VOLCANIC ROCKS
79.546. Ka-2, layer 5.
Fine grained, grey to black aphanitic rock, vesicular, in
thin section, an intermediate to basic volcanic rock showing
tablets of plagioclase with concentric composition zoning at
the edges,
the cores of which are eroded and with many
inclusions. Smaller laths of feldspar are set in a turbid,
possibly glassy groundmass making up 10 to 20% of the section.
79.43a. Ka-2, layer 5.
Fine grained, green-grey rock, feldspar and dark mineral
grains visible but not determinable in hand specimen. Weathers
to light grey. In thin section, ragged patches of opaque
minerals are present. Plagioclase feldspars fall into two size
ranges, one set 0.3 mm long and the other set 0.05 mm. The rock
is holo crystalline and appears as an intermediate to basic
volcanic rock.
79.43b. Ka-2, layer 5.
Dark grey aphanitic rock, feldspar laths to about 1 mm
long, dark blebs to about 0.25 mm visible. Weathers to light
grey. Thin section shows two sets of plagioclase laths at 0.3mm
and 0.05 mm long set in a groundmass of dusty, finely
crystalline material difficult to resolve. Rock appears to be an
-140altered volcanic of basic composition
79.549. Ka-2, layer 2.
Grey-brown rock showing yellow-green circular inclusions
up to 2 mm, fine laths of feldspar visible. In thin section, an
interlaced mesh of plagioclase crystals packed interstitially
with dusty and opaque weathering products can be seen. A volcanic
origin is probable.
79.550. Ka-2, layer 7.
Dense, black, aphanitic rock containing sparse, randomly
oriented feldspar laths and light coloured, sub-equant grains.
Weathers to an orange-brown friable material. In thin section
the rock appears to be an indurated crystal tuff.
79.551. Ka-2, layer 7.
Dark grey to black, mottled, aphanitic rock with sparse,
minute vesicles. In thin section, augite phenocrysts up to 0.3
mm diameter are set in a noncrystalline mesh of very fine
feldspar laths about 0.05 mm long. Opaque minerals about 3 to
5%. A basic volcanic rock.
79.553. Ka-2, layer 1.
Compact, dark grey to black rock, finely granular texture
with some lineation visible and small pockets of yellow
alteration products. In thin section, an intermediate to basic
volcanic rock showing a very fine grained, holocrystalline
texture, with some orientation of the small (about 0.06 mm)
feldspar laths, which show a marked structure with interstitial
tabular and equant crystals of hornblende.
79.555. Ka-3, spit 21.
Dark greenish-grey, compact rock, showing small (0.25 mm)
blebs of black material, weathering to brown-grey material. In
thin section, an intermediate to basic volcanic rock showing
tabular plagioclase up to 0.6 mm long set in very fine grained,
turbid matrix, with abundant small (0.04 mm) feldspar tablets.
79.294. Washed up on Werua Islet in 1980.
Grey-black, compact rock, weathering to pale reddish-grey.
Finely vesicular texture, groundmass aphanitic, small
phenocrysts of black pyroxene or amphibole visible. In thin
section, an intermediate to basic volcanic rock containing
abundant tabular plagioclase with concentrically zoned clear
overgrowths up to 0.6 mm in diameter, accompanied by very pale
green hornblende phenocrysts up to 0.4 mm, all set in a holocrystalline groundmass of small tabular feldspar crystals and
minute opaque grains.
79.295. Surface find.
Dark blue-grey, compact rock, fine grained texture,
indistinct lineation marked by dark material. In thin section,
an altered volcanic rock showing highly oriented, essentially
sub-parallel feldspar laths up to 0.3 mm long and abundant
interstitial holocrystalline material containing many minute,
euhedral prisms of pyroxene. Opaque minerals comprise about 10%
of the slide.
-14179.44. Ka-2, layer 4.
A dark brownish-black, compact, vesicular rock. Matrix
fine grained with sparse phenocrysts of pyroxene up to 1 mm long.
In thin section, this coarse grained basic volcanic rock largely
comprises a coarse intergrowth of plagioclase feldspar and
pyroxene.
-142-
APPENDIX 10
THE HUMAN BIOLOGY OF KAPINGAMARANGI
Phil Houghton
Some useful data on the living population of Kapingamarangi
has been provided by Schlaginhaufen(1929), and Emory(1965:64-79).
Emory's data relates to measurements and observations made by
Buck in 1947 on 59 adult males(Buck,1950:4), whilst Schlaginhaufen
studied some 34 adult males in 1908. Schlaginhaufen obtained a
mean stature of 1711 mm, and Buck 1723 mm, with a sitting height
ratio of 53.2. These are tall people in comparison with
neighbouring Micronesian and Melanesian islands(see Table 23),
and in this regard are characteristically Polynesian. The high
sitting height ratio is again Polynesian, showing relatively
long trunk, and short lower limbs.
The other traditional anthropometric measurements obtained
by these workers, and the various indices derived therefrom,
have little biological meaning, and it would not aid clarification
of biological relationships within Oceania to repeat them.
Schlaginhaufen collected hair samples, and also made
general comments on hair form amongst the population. The women
showed straight hair, as did the children, with men showing a
tendency to strongly waved hair; and he comments that truly
curly hair, that is the corkscrew type, was not seen on
Kapingamarangi. Again, these comments are in harmony with the
suggested Polynesian affiliations of the islanders, though of
course such a description of hair form could emerge from more
northern Micronesian islands as well.
Photographs of 14 individuals are supplemented by two
group photographs depicting a further 20 people in Schlaginhaufen's
work(1929). With broad lips and noses, long flat-sided faces
with vertical profile and mandibular prognathism revealing the
characteristic Polynesian form of jaw growth, these people may
be safely categorised as being of predominantly Polynesian
stock.
Miller studied the blood groups of 87 males and 82 females,
all believed to be of pure Kapingamarangi ancestry, during the
1950 expedition(Miller, 1953). His data are presented in Table 24.
The absence of group B supports the Polynesian relationships.
The prehistoric human material obtained from the present
Kapingamarangi excavations is small in quantity. Some six
individuals are represented, mostly by a few fragments of long
bones, but with rather more extensive remains of a mature
adult male, and jaw fragments of a six-year old. Fortunately,
these few pieces include some of the more informative parts of
the skeleton from the point of view of biological relationships
within Oceania.
-143The mature adult male shows the characteristic Polynesian
rocker jaw, from which it can be deduced that his general cranial
morphology was Polynesian. Femoral shafts from four people are
present, and all show the platymeric flattening characteristic
of Polynesians. The average platymeric index for these four is
73, which compares with Eastern Polynesian values of 65.2 for the
New Zealand Maori, 71.6 for Hawaiians, and a value of 77.5 from
Taumako, a Polynesian outlier in the Solomons. In general,
Homo sapiens
shows an index of over 80. While the biological
basis for this distinctive Polynesian femoral feature has not
yet been determined, it may be assumed that the lower limb
morphology in general for this group was Polynesian.
A permanent upper first molar from the child shows a
Carabelli cusp and a hypoconulid, and it has a crown area of 131.
However, no reliable conclusions concerning relationships can be
drawn from a single tooth.
None of the long bone fragments showed evidence of yaws,
a condition known to have been rife on at least one Polynesian
outlier, Taumako(Houghton, n . d . ) , and the skull fragments of the
adult male showed no bony evidence of anaemia, in the form of
spongy hyperostosis or cribra orbitalia. A second metacarpal from
this man had a cor ical thickness of 5.4mm, suggesting a wellnourished individual. The remaining skeletal fragments were
unsuited to radiological examination.
TABLE 23: AVERAGE STATURE FIGURES(cm) FOR KAPINGAMARANGI AND
VARIOUS OTHER OCEANIC GROUPS(Extracted from Shapiro,1933)
POLYNESIAN OUTLIERS
Kapingamarangi
Nukuoro
Taumako
Takuu
Nukumanu
Ontong Java
171.1
171.4*
175.lt
165.7
163.9
163.7
EAST MICRONESIA
Kiripati
Marshall Islands
169*. 5
163.6
POLYNESIANS
Tonga
Samoa
Society Islands
Fiji
Maori
Hawaii
173.0
171.7
171.4
170.9
170.6
169.5
ATOLL POLYNESIA
Funafuti
167.1§
-144Table 23 Continued
WEST AND CENTRAL MICRONESIA
West Carolines
Woleai
Ponape
Kusaie
Mortlocks
Truk
Yap
Palau
163.8
162.0
163.2
162.3
164.4
160.2
163.5
161.7
MELANESIANS
Santa Cruz
St Matthias
Tanna
New Caledonia
160.3
160.1
164.5
166.4
NB: * taken from Schlagenhaufen(1929:224)
t from Houghton(n.d.)
§ from Hedley(1897:235)
TABLE 24: BLOOD GROUP ANALYSIS OF KAPINGAMARANGI
(after Miller, 1950:41-42)
AB
Males
Females
Totals
A
40
32
B
72
-
-
0
47
50
97
Total
87
82
169
-145-
REFERENCES
Anell,B.
1955
Contribution to the history of fishing in the
southern seas. Studia Ethnographica Upsaliensia 9
Anon.
1944
Kapingamarangi. Joint Intelligence Centre, Pacific
Ocean Areas Information Bulletin(JICPOA) 26-44.
1972
Ponape Almanac. Instructional
n .d,
Pilot chart of the Pacific ocean.#1401W,#2601W,
Anon
Serves
Centre,Ponape
Anon,
U.S.Hydrological
Baker,R.H.
1951
Office.
The avifauna of Micronesia, its origin,evolution,
and distribution. University
of Kansas
Museum of Natural History 3(l):l-359.
Bayard,D.T,
1976
The Cultural
Relationships
of
the
Publications>
Polynesian
Outliers.
University of Otago, Studies in
Prehistoric Anthropology 9.
Beasley,H.G.
1928
Pacific
Island
Records:
Fish
hooks.
Seeley,Service
and Co., London.
Biggs,B.
1979
Proto-Polynesian word list I I . U n i v e r s i t y
Working Papers in
of
Auckland
Anthropology.
Buck,P.H.
1950
Material
Culture
of Kapingamarangi.
B.P.Bishop
Museum Bulletin 200.
Carroll,V.
1965
An outline of the structure of the language of
Nukuoro. Journal
74(2) :192-226.
Carroll,V.(Ed.)
Adoption
1970
of
in Eastern
the Polynesian
Society
Oceania.ASAO Monograph 1.
University Press of Hawaii.
Cernohorsky,W.O.
1971
Marine shells
Sydney.
of the Pacific
J.Pacific Publications,
Cernohorsky,W.0.
1972
Marine shells
Sydney.
of the Pacific
JJ.Pacific Publications,
-146Davidson,J.M.
1967
An archaeological assemblage of simple fish-hooks
from Nukuoro atoll. Journal
of the
Polynesian
Society
76 (2) :177-196.
Davidson,J.M.
1971
Archaeology
on Nukuoro atoll:
a Polynesian
outlier
in the Eastern
Caroline
islands.
Auckland Institute
and Museum Bulletin 9.
Eilers,A.
1934
Elbert,S.H.
1949
Inseln
um Ponape.
Ergebnisse der Sudsee-Expedition
1908-1910. Friederichsen, de Gruyter & Co,Hamburg.
Uta-Matua and other tales of Kapingamarangi.
Journal
Emory,K.P.
1965
of
American
Kapingamarangi:
Polynesian
atoll.
Folklore.
62 (245) :240-246.
social
and religious
life
of a
B.P.Bishop Museum Bulletin 228.
Fosberg,F.R.
1946
Botantical Report on Micronesia. Economic
of Micronesia
13. U.S.Commercial
Co.
Survey
Fujimura,K. and Alkire,W.H.
1979
Archaeological test excavations on Faraulep, Woleai,
and Lamotrek in the Caroline Islands of Micronesia.
Unpublished
Manuscript.
Gantt,P.A.
1946
Livestock survey of Micronesia. Economic
of Micronesia
IS. U.S.Commercial
Co.
Gosline,W.A. and Brock,V.E.
1960
Handbook of Hawaiian
Hawaii.
Green,R.C.
1970
Hambruch,P.
1914
Hedley,C.
1897
Fishes.
Survey
University Press of
Settlement pattern archaeology in Polynesia,
pp 13-32 In: Green,R.C. and Kelly,M.(Eds.)
Studies in Oceanic Culture History 1.
Pacific
Anthropological
Records
11.
Nauru. Ergebnisse der Sudsee-Expedition,1908-1910.
Friederichsen, Hamburg.
The Ethnology
Memoir III.
of
Funafuti.
Australian Museum
-147Houghton,P.
n.d.
The people of Namu. In: Leach,B.F. and Davidson,
J.M.(Eds.) q.v.
Kennedy,D.G.
1931
Field
notes
on the culture
of Vaitupu,
Ellice
Memoirs of the Polynesian Society, Volume 9.
Koch,G.
1961
Kubary,J.S.
1900
Die Matevielle
Kultur
Volkerkunde, Berlin.
der
Ellice-InseIn.Museum
Islands.
fur
Beitrag zur Kenntnis der Nukuoro - oder Monterverde
-Inseln. Mitteilungen
der Geographische
Gesellschaft
16:71-138.
Lampert,R.J.
1965
An archaeological investigation of Te Aka village,
Ocean island: a preliminary report.
Unpublished
Manuscript.
Leach,B,F.
1979
Excavations in the Washpool valley, Palliser Bay.
pp 67-136 In: Leach,B.F. and Leach,H.M.(Eds.)
Prehistoric
man in Palliser
Bay. National Museum
Bulletin 21.
Leach,B.F. and Anderson,A.J.
1979a The role of labrid fish in prehistoric economics
in New Zealand. Journal
of Archaeological
Science
6:1-15.
Leach,B.F. and Anderson,A.J.
1979b Prehistoric exploitation of crayfish in New Zealand,
pp 141-164 In: Anderson,A.J.(Ed.) Birds of a
Feather.
British Archaeological Reports S62.
Leach,B.F. and Davidson,J.M.
1977
Fishing methods and seasonality at Paremata(N160/50).
20(3):
N.Z.Archaeological
Association
Newsletter
166-175.
Leach,B.F. and Davidson,J.M.
n.d.
Taumako: a Polynesian
outlier
in the
Eastern
Solomon Islands . Unpublished Manuscript.
Levison,M., Ward,R.G., and Webb,J.W.
1973
The settlement
of Polynesia:
a computer
Australian National University Press.
Lewis,D.
1972
Vie, the
Press.
navigators.
simulation.
Australian National University
-148Lieber,M.D.
1968a
Lieber,M.D.
1968b
The nature of the relationships between kinship
and land tenure on Kapingamarangi atoll.
Unpublished
PhD thesis.
University
of
Pittsburgh.
Porakiet:
a Kapingamarangi
colony
on Ponape.
Eugene, University of Oregon, Department of
Anthropology.
Lieber, M.D.
1970
Adoption on Kapingamarangi. pp 158-205 In:
Carroll, 1970 q.v.
L i e b e r , M.D.
1974
Land t e n u r e on1 Kapingamarangi. pp 70-99 I n :
Lundsgaarde , 1 9 74 q . v .
Lieber,M.D.(Ed.)
1977a Exiles
and migrants
of Hawaii.
Lieber,M.D.
1977b
Lieber,M.D.
1979
in
Oceania.
University Press
The process of change in two Kapingamarangi
communities, pp 35-67 In: Lieber,1977a q.v.
Kapingamarangi land history. Unpublished
for Ponape Historic
Preservation
Committee.
report
196pp.
Lieber,M.D. and Dikepa,K.H.
1974
Kapingamarangi
lexicon.
University Press of
Oceania.
University Press of Hawaii.
Hawaii.
Lundsgaarde,H.P.(Ed.)
1974
Land tenure
McKee,E.D.
1956
in
Geology of Kapingamarangi
atoll,
Atoll Research Bulletin 50.
Merrill,E.D.
1946
Plant
life
New York.
of
Miller, R.E.
1953
Health
report
Bulletin 20.
the
Pacific
Caroline
islands
.
worZd.MacMillan Co.,
of Kapingamarangi.
Atoll Research
Munro,I.S.R.
1967
The fishes
of New Guinea.
Department of Agriculture,
Stock and Fisheries, Port Moresby, New Guinea.
Niering,W.A.
1956
Bioeoology
terrestrial
of Kapingamarangi
atoll,
Caroline
islands
aspects.
Atoll Research Bulletin 49.
-149Parham,B.E.V.
1972
Plants
Pawley,A.
1967
of
Samoa.
DSIR Information Series 85
The relationships of Polynesian outlier languages
Journal
of the Polynesian
Society
76(3):259-296.
Pawley,A. and Green,K.
1971
Lexical evidence for the proto-Polynesian
homeland. Te Reo 14:1-35.
Ray,S.H.
1917
The people of Greenwich atoll,Western Pacific
Ocean. Man 17 (130) :187-190.
Riesenberg,S.H. and Gayton,A.H.
1952
Caroline island belt weaving. Southwestern
of Anthropology.
8:342-375.
Journal
Schlaginhaufen,0.
1929
Zur Anthropologie der mikronesischen Inselgruppe
Kapingamarangi(Greenwich-Inseln).Arohiv der
Julius
Klaus-Stiftung
fur
Vererbungsforschung,
Sozialanthropologie
und Rassenhygiene
4(3):219-287.
(Translation by K.J.Dennison).
Shapiro,H.L,
The physical
characteristics
of the
Ontong
1933
Javanese.
Anthropological papers of the
American Museum of Natural History.23(3):225-278
Thilenius,G,
Ethnographische
Ergebnisse
aus
Melanesien.
1903
Reisebericht.
Die Polynesischen
Inseln
an der
Ostgrenze
Melanesiens.
Kaiserliche LeopoldinischCarolinische Deutsche Akademie der Naturforscher 80
Thomas,W.L,
1965
Titcomb,M.
1972
Vokes,H.E.
1980
The variety of physical environments among Pacific
islands, pp 7-38 In: Fosberg,F.R.(Ed.) Man's
Place
in the island
ecosystem.
Bishop Museum Press.
Native
use of fish
of Hawaii.
in
Hawaii.
University Press
Genera of the bivalvia.
Paleontological Research
Institution, Ithica, N.Y.
Walsh,D.S. and Biggs,B.
1966
Proto-Polynesian
word
list
I.
Te Reo Monographs.
-150Ward,R.G., Webb,J.W., and Levison,M.
1973
The settlement of the Polynesian outliers: a
computer simulation. Journal
of the
Polynesian
Society.
82 (4) :330-342.
Whitmore,T.C.
1966
Guide to
Islands.
Wiens, H.J.
1956
Wilhm,J.L.
1968
the forests
of the British
Oxford University Press.
Solomon
The geography
of Rapingamarangi
atoll
in
.Carolines.
Atoll Research Bulletin 48.
the
Use of biomass units in Shannon's formula.
Ecology.
49:153-156.
eastern