CHAPTER ONE

INTRODUCTION
by Patrick V. Kirch

A    s INMANYPARTS oFrmwoRLD, rockshelters have

been a favored site tpe for archaeological studies
in the Hawaiian Islands. J.F.G. Stokes' pioneering
excavations in 1913 on Kaho'olawe Island (Kirch

1985:12) penetrated the deposits of a fisherman's shelter
rich in well-preserved organic materials. In 1950, when
Hawaiian archaeology was rejuvenated by K. P. Emory,
the Kuli'ou'ou Shelter on O'ahu Island provided the

setting (Emory and Sinoto 1961). Over the past three
and one-half decades, more than 50 rockshelters have

been excavated or tested, on every major island (see, as
examples, Bonk 1954; Chapman and Kirch 1979;

Emory and Sinoto 1961; Kirch 1979b; McCoy 1977;
Soehren 1966). However, almost all of these sites are
situated within the coastal zone, less than 1 km from
the shoreline. Indeed, many of these shelters were

selected for excavation because they would predictably
yield fishing gear, especially fishhooks, so important
for Hawaiian relative chronology (Emory, Bonk, and
Sinoto 1959). Only the small shelter sites within the
Mauna Kea adz quarry, at 3,500 m on Hawai'i Island,
are exceptions to this pattern (McCoy 1977).

This monograph presents the results of excavations
in three rockshelter sites located in a wholly different

environmental setting: an interior valley locale some 6
km from the coast, on the island of O'ahu. Excavations
in the Anahulu Valley rockshelters thus provide an

important test case for the putative general Hawaiian
pattern of early coastal settlement, followed by later
"inland expansion" of populations (Hommon 1976,

1986; Kirch 1985:296-7). In the stratigraphic sequences

of the three rockshelters reported herein, it is possible to
track the initial stages of exploitation of interior valley
resources, followed by more intensive utlization and
permanent occupation immediately prior to European
contact.

THE ANAHULU VALLEY PROJECT
The 1982 Anahulu Valley Project, organized

jointly by the author and Marshall Sahlins (University
of Chicago), is the shortened designation for a research
endeavor entitled "The Archaeology of Ethnohistory:

The Historical Transfonnation of Hawaiian Society and
Economy" (National Science Foundation Grant No.
BNS 82-05621). The Project built upon extensive
edtnohistorical research on the early Hawaiian

Kingdom, carried out by Sahlins from 1971 to 1976,
and on a pilot archaeological study in the Anahulu

Valley by Kirch in 1974 and 1976 (Kirch 1979a). The
aim of the 1982 Project was an inter-disciplinary

archaeological and ethnohistorical investigation of

socio-political organization, economy, land use, and
settlement pattern in the Anahulu Valley, situated on
northwest O'ahu Island (fig. 1.1). As a land unit

(ahupua'a) within the political distict of Waialua,
Anahulu had come under the control of several

important high chiefs of the early Hawaiian Kingdom
from AD. 1804 and into the mid-nineteenth centuy.
The consequent wealth of archival and ethnohistoric
material relating to Anahulu was matched by an

undisturbed archaeological landscape in the middle and

Anahdul

KAHUU PT. 100

1;~~~~6

t>MALAEKAIHANA 0'

I

KONA

Figure 1.1. O'ahu Island, showing the location of Waialua District and the Anahulu Valley (after Kirch 1985, fig. 86).

upper reaches of the valley. Thus, the expectation was
that the transformation of indigenous Hawaiian society
and economy in the early decades following European
contact and during the formation of the Hawaiian

Kingdom, might be elucidated through the analysis of
this particular valley system (Kirch, in press).

The 1982 fieldwork in Anahulu Valley concentrated
on three main types of archaeological site: (1) stone-
faced pondfield irrigation systems constructed on

alluvial terraces in the valley bottom; (2) a series of
open habitation sites on the colluvial slopes; and (3)
three rockshelters with occupation deposits, in the

middle valley. The first two site types proved to date

almost wholly to the nineteenth century, and to reflect a
phase of permanent occupation and intensive

agricultural utilization of the interior Anahulu Valley

following the conquest and occupation of O'ahu by the
forces of Kamehameha I in 1804. The rockshelters, in

contrast, yielded a sequence of prehistoric use extending
back to the thirteenth century A.D., with only limited
occupation in the very early historic period.

While the Anahulu rockshelter sites are thus of

considerable interest for Hawaiian prehistory, they bear
less directly on the principal aim of the Anahulu Valley
Project, namely the analysis of historic-period

transformation of Hawaiian economy and society.

Nonetheless, because the shelters yielded significant

archaeological materials, and provided some of the few
examples of truly interior valley rockshelter sites in the
entire archipelago, it was important that the results of
our excavations in these sites be presented in full.

Hence, the decision to publish this monograph on the
rockshelter excavations, separately from the combined

report on our ethnohistorical and archaeological study of
the valley from 1804 to the later nineteenth centry
(Sahlins and Kirch, in prep.).

ROCKSHELTER SITES OF THE

ANAHULU VALLEY

MTe geomorphological characteristics of Hawaiian

valleys vary substantially depending upon their position

2

Introduction

on either windward or leeward slopes of the volcanic

mountain ranges. Windward valleys, subject to heavy

rainfall and consequent sheet erosion, mass wasting, and
rapid chemical weathering of rock, are typically

amphitheatre-headed, with U-shaped profiles (Wentworth
1928; MacDonald and Abbot 1970:172-9). Leeward

valleys, especially in their lower elevations farther from
the heavy rainfall zone at the mountain crest, tend to be
more linear, narrow, and deeply incised, with

streamflow the main erosive agent. This is the case

with Anahulu and other valleys (e.g., Waimea, Opaeula,
Helemano) that incise the leeward, arid southwesterly

slopes of the Ko'olau Mountain range on O'ahu Island.
Consequendy, for the first 6 km inland from the coast,

the valley walls consist largely of vertical cliffs (cutting
through alternating flows of aa and pahoehoe lava) and
talus scree slopes. (Farther inland than 6 km, the

rainfall gradient becomes sufficiently steep to change
the erosional regime, and the valley profile becomes

more U-shaped, lacking cliffs.) These cliffs provide an
ideal geological setting for rockshelters, which were
formed where the stream undercut massive aa lava
formations by eroding the less resistant underlying

clinker layers. Many of these stream-cut notches are

now stranded well above the present stream level due to
continued downcutting of the Anahulu Stream.

The first rockshelter to be archaeologically

investigated in the Anahulu Valley was Site OA-D6-
14* (originally designated Site 06), a small shelter at
the base of a vertical cliff on the north side of the

valley. Site D6-14 was excavated in the late 1950s by
an amateur, Mr. A. Andersen, with advice from Dr.

Kenneth P. Emory of the Bishop Museum. The site

evidently spanned both the prehistoric and early historic
periods, although no radiocarbon dates were ever

obtained. A brief note on the site was published by

Kirch (1979a:61-2) based on Andersen's field notes in
the Bishop Museum.

With the commencement of a pilot archaeological
study in the Anahulu Valley in 1974 (Kirch 1979a),
reconnaissance surveys were made along the valley

walls for additional rockshelters that might contain well
stratified occupation deposits. Five such shelters were
recorded (D6-52, -56, -57, -58, and -60) (Kirch
1979a:59-60), and Site D6-52 was selected for

investigation. This shelter lies just above the stream
flats (irrigated for taro cultivation during the historic
period) on the north side of the valley, about 4 km

inland. The 1974 test excavations showed that the site

was well stratified, yielding the first occupation

sequence for the middle Anahulu Valley. Consequently,
additional excavations were carried out in 1976, with a
total area of 6.21 m2 being exposed in the two field
seasons (fig. 1.2).

Radiocarbon dates on samples of charcoal and ash
from two hearths in Site D6-52 yielded ages of 185 ?
80 and 325 ? 80 B.P. The deposits yielded a range of
faunal and carbonized floral materials, as well as 153
portable artifacts. Analyses of these materials are

presented in detail in Kirch (1979a:32-50). The evidence
from the D6-52 rockshelter excavation was interpreted
as follows:

The late prehistoric utilization of middle
Anahulu Valley was probably of a transient

nature, as a resource zone or area exploited by a
permanent, coastal-dwelling population. Rock-
shelters such as that at Keae or Site 06 served
as temporary bases-occupied for a few days or
perhaps weeks at a time-for local groups

practicing shifting cultivation, collecting forest
materials such as firewood or cordage fibers,

catching birds, or collecting freshwater shell-

fish and shrimp from the stream (1979a:51-2).

RESEARCH DESIGN

With the decision in 1982 to expand the Anahulu
pilot research into a major coordinated ethnohistoric-
archaeological program, it was clear that further
excavations in mid-valley rockshelters would be

essential to test the model outlined above. Our initial

plans called for the excavation of two sites, D6-58 and
D6-60, both large overhang shelters on north and south
sides of the valley 5-6 km inland from the coast. After
work had commenced at these sites, a third, smaller

shelter (D6-36) was discovered not far from Site D6-58,
and was included in the excavation program.

The 1982 rockshelter excavations comprised one
facet of a larger research program focussed on the
historical transformation of indigenous Hawaiian

society and economy (see above). Within this research
framework, the rockshelters were regarded as the most

likely site class to yield materials extending back prior
to the European contact period, thus providing the

evidential basis for a prehistoric occupation sequence
from the middle valley. Understanding the patterns of

prehistoric occupation and utilization of the valley was

*The site numbering system used in this monograph is that of the Bernice P. Bishop Museum. The prefix OA-
refers to O'ahu Island, and for brevity has been dropped from the numbers henceforth. D refers to the district of

Waialua, and 6 to the ahupua'a or land unit of Kawailoa. The final number is the individual site designation within the
ahupua'a unit.

3

Anahiu

Figure 1.2. 1976 excavation in the main occupation floor in Site D6-52, a rockshelter in the Keae area of Anahulu
Valley.

obviously essential if the rapid and intense changes of

the early European-contact period were to be isolated. At
the commencement of fieldwork in 1982, it was unclear
whether the middle and upper portions of the Anahulu

Valley had been permanently occupied in prehistory, or
whether the extensive taro pondfield irrigation systems
of the valley flats had been in use prior to European

contact. The results of four months of excavations, not
only in the three rockshelters, but also in nine open

habitation sites, and several irrigation systems, were to
show that the irrigation systems and open, permanent
residential complexes were developments of the early
contact period (Kirch, in press; Sahlins and Kirch, in
prep.). Thus, it is indeed the rockshelters that hold the
key to understanding the prehistoric period in the
interior portions of the Anahulu Valley.

Given the overall research framework of the 1982
Anahulu Valley Project, the rockshelter excavations

were oriented primarily at obtaining a temporal sequence
of occupation. Our field strategy thus focussed on

vertical, stratigraphic exposures in several shelters,

rather than extensive horizontal excavations of shelter

floors. However, in Site D6-60 we did open up an area
of 18 m2 in an effort to gain some understanding of the
horizontal distributions of features, artifacts, and faunal
materials.

At the time our excavations were undertaken, there
was increasing awareness that prehistoric Polynesian
occupation of the Hawaiian Islands had resulted in
significant environmental changes to this formerly

isolated, and vulnerable island ecosystem (Kirch 1982).
Along with several other Hawaiian archaeological

projects at this time (e.g., Clark and Kirch 1983; Schilt
1984), considerable emphasis was placed on recovering
a full range of sedimentological, faunal, and floral
materials from the rockshelters, with the aim of

4

Introduction

reconstructing the local sequence of ecological changes
precipitated by prehistoric human use of the Anahulu
Valley. In excavation procedure, this objective was

reflected in the use of fine screening (both 1/8 and 1/4
inch mesh sieves) of all deposits, and in the use of

column samples for extraction of minute seeds and land
snails. As the various chapters in this monograph
attest, considerable effort was expended on the

laboratory analysis of sediments, non-marine molluscs,
charcoal, botanical remains, and vertebrate and inverte-
brate faunal assemblages, in addition to artifactual

materials. In many respects, these analytical efforts were
pioneering, and our results at times reflect the

limitations of inadequate reference collections (as with
the charcoal identification project, see chapter 7), or
unanticipated problems of sample size (as in the

analysis of flotation samples of seeds, see chapter 6).

Nonetheless, the results reported in the various contri-
butions to this volume reveal the potential of carefully
controlled rockshelter excavations to yield a wealth of
information not only on Hawaiian cultural prehistory,
but on biogeographic and ecological change as well.

The research design for the investigation of the

Anahulu rockshelters was thus oriented by four major
objectives: (1) To establish a chronology for the

prehistoric utilization and occupation of the middle

valley region, based on radiocarbon dating. (2) To test
the model derived from the 1974-76 excavation at

rockshelter Site D6-52, that the shelters were temporary
occupations, part of a pattern of low-intensity exploita-
tion of the interior valley during the prehistoric period.
(3) To determine whether the occupation sequences of
any of the shelters exhibited a shift from temporary to
permanent use in the late prehistoric or early historic
periods, that might correspond with a more intensive
exploitation of the upper valley. Objectives 2 and 3

required the acquisition of stratigraphically-controlled

assemblages of artifacts and faunal materials, as well as
of non-portable cultural features. (4) To track the

sequence of environmental conditions in the vicinity of
the rockshelters throughout the period of their human
utilization, as reflected in the rockshelter sediments,
terrestrial molluscs, archaeobotanical materials, and
charcoal.

ENVIRONMENTAL CONTEXT OF THE

MIDDLE ANAHULU VALLEY
The three rockshelter sites reported in this

monograph are located in what we have termed "middle"
Anahulu Valley, a distance of approximately 6 km from
the coast (fig. 1.3). In this stretch, the valley floor is

only about 150 km wide, flanked on either side by steep
talus slopes and vertical cliffs. The valley floor has an

elevation of between 98-110 m-above sea level, while
the adjacent tablelands above the cliffs are about 215 m
above sea level. The cliffs, into the bases of which the
rockshelters are cut, expose sections through the
basaltic lavas of the Ko'olau Volcanic Series, of
Pliocene age (Stearns and Vaksvik 1935).

The Anahulu Stream meanders across the valley,
dividing the floor into a series of alluvial flats or
terraces, alternatively to the south or north of the
steam. These alluvial teraces were the focus of
intensive pondfield irrigation of taro (Colocasia

esculenta) in the early historic period. Prior to terracing
for pondfields, the alluvial flats would have been ideal
environments for shifting cultivation. The soils on
these terraces consist of varying facies of the

Kawaihapai Soil Series, well-drained mollisols (Foote
et al. 1972:63).

The valley has a steep rainfall gradient (Rosenau,

Lubke, and Nakahara 197 1:D7-13, fig. 6), with the area
in which the rockshelter sites are located receiving
approximately 1,250 mm annually. The Anahulu

headwaters are inundated with as much as 7,500 mm
annually, ensuring permanent streamflow even in the

lower elevation, more arid parts of the valley. Since the
beginning of this century, however, much of this

streanflow has been diverted out of the valley to the

adjacent tablelands for sugar cane irrigation. Thus today
the Anahulu Stream in the vicinity of the rockshelter
sites exhibits intermittent flow. Prehistorically,

however, there is no doubt that the Anahulu Stream

flowed permanently throughout the length of the valley.

The flora and vegetation pattns of the middle
Anahulu Valley reflect several centuies of human-

induced ecological change, beginning with prehistoric
exploitation and continuing in the historic period with
ranching and plantation agriculture. The flora of the

valley is dominated by exotic introductions, with only
limited numbers of indigenous or endemic species that
hint at the pre-human phytogeography. The middle

valley falls witiin Zone C of Ripperton and Hosaka's
(1942) classification, with the valley slopes and side

walls dominated by the exotics Leucaena leucocephala,
Schinus terebinthifolius, and Lantana camara. (This is
also comparable to the "Guava zone" of Hosaka 1937.)
Psidium guayava is also common on the lower slopes

and alluvial flats. The alluvial flats are shaded by stands
of large Mangifera idi ca, although Eugenia cwninii is
also common in places. Among the few indigenous-

endemic species found on the valley walls are isolated

Acacia koa and Canthium odoratwn trees, and shrubs of
Styphelia tameiameiae, and on the talus slopes,
Erythrina sandwicensis.

There are a number of feral cultigens in the middle
valley, which have evidently naturalized following their

5

Anahiu

cis

c)

CJ

Co
-

0c

4-

4)

'0

cis

'U

CI

.4

(U

I-

C)

4)

0

CO

U

C-

0

CC

C
Co

.4)

CU

on

*CC
o)
4)

Co

on N

* _

6

Introduction

introduction during the period of Polynesian utlization
of the middle valley. Among these are the breadfruit

(Artocarpus altilis), of which scattered trees are found at
the base of the talus slopes, the ti plant (Cordyline

fruticosa) which is common on the lower slopes and

alluvial flats, the bitter yam (Dioscorea bulbfera) found
on the colluvial slopes, and noni (Morinda citrifolia)
which grows commonly on the alluvial flats.

The fauna of the valley has also changed radically
in historic times. Native land birds, such as the

honeycreepers, are now either extinct or confined to the
highest portions of the Ko'olau Range, well inland of
the rockshelter zone. The same is true of most of the
endemic land snails, such as the brightly-colored
Achatinella spp. tree snails, which are currently

threatened with extinction. The native goboid fishes

(Chonophorus spp.) and shrimp, along with the edible

freshwater snail (Neritina granosa), that formerly would
have been common in the Anahulu Stream have been
virtually eliminated by streamflow alteration and

pollution (especially by fertilizers and insecticides) from
plantation agriculture. The valley's fauna is dominated
by exotic introductions, including a number of land
birds (mynahs, rice birds, house finches, etc.), and
mongoose.

REFERENCES CITED

Bonk, W.J. 1954. "Archaeological Excavadons on West

Molokai," M.A. Thesis, University of Hawaii.

Chapman, P.S., and P.V. Kirch. 1979. Archaeological

Excavations at Seven Sites, Southeast Maui,

Hawaiian Islands. Dept. of Anthropology Report
79-1. B.P. Bishop MuseunL Honolulu.

Clark, J., and P.V. Kirch, eds. 1983. Archaeological

Investigations in the Mudlane-Waimea-Kawaihae
Road Coridor, Island of Hawaii: An

Interdisciplinary Study of an Environnental

Transect. Dept. of Anthropology Report 83-1. B.P.
Bishop Museum. Honolulu.

Emory, K.P., W.J. Bonk, and Y.H. Sinoto. 1959.

Hawaiian Archaeology: Fishhooks. B.P. Bishop
Museum Special Publication 47. Honolulu.
Emory, K.P., and Y.H. Sinoto. 1961. Hawaiian

Archaeology: Oahu Excavations. B.P. Bishop
Museum Special Publication 49. Honolulu.

Foote, D.E. et al. 1972. Soil Survey of the Islands of

Kauai, Oahu, Maui, Molokai, and Lanai, State of

Hawaii. U.S. Dept. Agriculture, Soil Conservation
Service. Government Printing Office. Washington.
Hommon, R.J. 1976. "The Formation of Primitive

States in Pre-Contact Hawaii." PhD. Dissertation,
Univ. of Arizona. (University Microfilms.)

. 1986. Social evolution in ancient Hawaii. In
P.V. Kirch, ed., Island Societies: Archaeological

Approaches to Evolution and Transformation, pp.
55-68. Cambridge University Press.

Hosaka, E.Y. 1937. Ecological and floristic studies in

Kipapa Gulch, Oahu. BP. Bishop Museum

Occasional Papers 13(17):175-232. Honolulu.

Kirch, P.V. 1979a. Late Prehistoric and Early Historic

Settlement-Subsistence Systems in the Anahulu

Valley, O'ahu. Dept. of Anthropology Report 79-2.
B.P. Bishop Museum, Honolulu.

. 1979b. Marine Exploitation in Prehistoric
Hawai'i: Archaeological Excavations at
Kalahuipua'a, Hawai'i Island. Pacific

Anthropological Records 29. Dept. of

Anthropology, B.P. Bishop MuseunL Honolulu.

____. 1982. The impact of the prehistoric Polynesians

on the Hawaiian ecosystenm Pacific Science
36(1):1-14.

. 1985. Feathered Gods and Fishhooks: An
Introduction to Hawauan Archaeology and

Prehistory. University of Hawaii Press. Honolulu.

in press. Production, intensification and the

early Hawaiian kingdonL In DE. Yen, ed., Pacific
Production Systems. Australian National
University.

MacDonald, G.A., and A.T. Abbot. 1970. Volcanoes in

the Sea: The Geology of Hawaii. University of
Hawaii Press. Honolulu.

McCoy, P.C. 1977. The Mauna Kea Adze Quarry Proj-

ect a summary of the 1975 field investigations.
Journal of the Polynesian Society 86(2):223-44.
Ripperton, J.C., and E.Y. Hosaka. 1942. Vegetation

Zones of Hawaii. Hawaii Agricultural Experiment
Station Bulletin 89. Honolulu.

Rosenau, J.C., E. Lubke, and R. Nakahara. 1971.

Water Resources of North-Central Oahu, Hawaii.
Geological Survey Water-Supply Paper 1899-D.
Government Printing Office. Washington.

Sahlins, M., and P.V. Kirch. in prep. Anahulu: The

Archaeology of History in the Early Sandwich
Islands Kingdom

Schilt, A. 1984. Subsistence and Conflict in Kona,

Hawaii. Dept. of Anthropology Report 84-1. B.P.
Bishop Museum. Honolulu.

Soehren, LJ. 1966. 'Hawaii Excavations, 1965."

Typescript in Dept of Anthropology, B.P. Bishop
Museum. Honolulu.

Stearns, H.T., and KN. Vaksvik. 1935. Geology and

Ground-Water Resources of the Island of Oahu,
Hawaii. Bulletin 1, Division of Hydrography,
Territory of Hawaii. Honolulu.

Wentworth, C.K. 1928. Principles of stream erosion in

Hawaii. Journal of Geology 36:385-410.

7