Contents lists available at ScienceDirect

Journal of Archaeological Science: Reports

journal homepage: www.elsevier.com/locate/jasrep

Volcanic glass at Kualoa, O‘ahu, Hawaiian Islands: Paired technological and

geochemical sourcing analyses of an expedient tool industry

Nathaniel J. DiVito

, Thomas S. Dye

⁎

, Kawena Elkington

, JoLynn Gunness

, Eric Hellebrand

Elaine Jourdane

, Steven Lundblad

, Nicole Mello

, Peter R. Mills

, John M. Sinton

Honua Consulting, 4348 Waialae Ave., #254, Honolulu, HI 96816, United States

60 North Beretania St., Apt. 3201, Honolulu, HI 96817, United States

Kamakakūokalani Center for Hawaiian Studies, Hawai‘inuiākea, University of Hawaii at Mānoa, 2645 Dole Street, Honolulu, HI 96822, United States

Honolulu, HI, United States

Department of Earth Sciences, University of Hawai‘iatMānoa, Honolulu, HI 96822, United States

Department of Geology, University of Hawai‘i at Hilo, 200 Kawili St., Hilo, HI 96720, United States

Department of Anthropology, University of Hawai‘i at Hilo, 200 Kawili St., Hilo, HI 96720, United States

ARTICLE INFO

Keywords:

Volcanic glass

Exchange

Hawai‘i

EDXRF

Microprobe

ABSTRACT

The results of paired technological and geochemical sourcing analyses of 1258 pieces of volcanic glass collected

during archaeological investigations in the coastal portion of Kualoa on the windward coast of O‘ahu, Hawaiian

Islands are reported.

Geochemical analyses by EDXRF and microprobe indicate most of the volcanic glass pieces recovered at

Kualoa derived from a source in the Wai‘anae Range on the leeward side of O‘ahu. Also present are three pieces

from Pu‘uwa‘awa‘a on Hawai‘i Island. Much of the rest of the material resembles sources in the Ko‘olau Range,

however this is a common geochemical makeup that is widely distributed on O‘ahu and other islands in Hawai‘i

and precise sources for this material have not been identiﬁed.

Technological analyses of volcanic glass nodules, cores, ﬂakes, and debris indicate that volcanic glass from

each of the sources was brought to Kualoa as unworked nodules, a ﬁnding that supports previous conclusions

that Hawaiians enjoyed direct access to volcanic glass sources. A reduction sequence based on observations of

broken and whole ﬂakes and of cortex on the dorsal faces of whole ﬂakes with and without edge damage

indicates that volcanic glass users habitually chose larger ﬂakes for tasks that resulted in macroscopic traces of

edge damage. Smaller ﬂakes were either used for tasks that did not leave macroscopic traces of edge damage or

were discarded without being used.

Paired technological and geochemical sourcing analyses indicate a volcanic glass industry at Kualoa oriented

to the production of ﬂakes that were likely used for a variety of expedient cutting and scraping tasks by people

with direct access to unworked material from local and non-local volcanic glass sources up to 300 km away. The

picture that emerges from the paired technological and geochemical sourcing analyses is of a highly connected

society in which people traveled widely and enjoyed support for their productive activities with unfettered

access to sources of volcanic glass.

1. Introduction

Recognition that pristine states developed rapidly in Hawai‘i

(Hommon, 2013; Kirch, 2010) following discovery and colonization of

the islands around

AD 1000–1100 (Athens et al., 2014) raises questions

about the extent to which political development aﬀected the quotidian

habits of commoners. Two lines of thought have developed. One views

the local ahupua‘a community as self-suﬃcient, self-contained, and

managed by a hierarchy of local oﬃcials appointed by the king. This

managerial view of political development contrasts with a laissez-faire

view of the ahupua‘a community as a tribute district where local oﬃ-

cials organized tribute collection and presentation, but were not

https://doi.org/10.1016/j.jasrep.2019.102117

⁎

Corresponding author.

E-mail addresses: divito.nathan@gmail.com (N.J. DiVito), tsd@tsdye.online (T.S. Dye), kjelking@hawaii.edu (K. Elkington), gunness@hawaii.edu (J. Gunness),

ericwgh@hawaii.edu (E. Hellebrand), ejourdane@gmail.com (E. Jourdane), slundbla@hawaii.edu (S. Lundblad), nmello@hawaii.edu (N. Mello),

millsp@hawaii.edu (P.R. Mills), sinton@hawaii.edu (J.M. Sinton).

Journal of Archaeological Science: Reports 30 (2020) 102117

concerned with the quotidian habits of commoners.

Recent progress in the geochemical characterization of stone tools

and identiﬁcation of source locations has produced results cited in

support of both views. The managerial view is supported by the dif-

ferential distribution of ﬁne-grained basalt imported from Hawai‘i

Island among housing compounds of diﬀerent size and complexity on

Maui Island interpreted as an example of political control over access to

high quality tools and their use within the ahupua‘a community (Kirch

et al., 2012). The laissez-faire view is supported by the distribution of

volcanic glass from the Pu‘uwa‘awa‘a source across political boundaries

on Hawai‘i Island interpreted as an example of direct access by people

from many ahupua‘a communities to an important source of high

quality raw material free of political control (Putzi et al., 2015; McCoy

et al., 2011).

This paper presents the results of paired technological and geo-

chemical sourcing analyses of 1258 volcanic glass pieces collected

during archaeological investigations carried out in the 1970’s and early

1980’s in the coastal portion of Kualoa Ahupua‘a, located in the

Ko‘olaupoko District of O‘ahu Island (Fig. 1). The technological ana-

lyses identify a reduction sequence designed to produce ﬂakes larger

than 8 × 7 mm in size. The geochemical sourcing results identify an

important O‘ahu Island volcanic glass source at Pu‘uKa‘īlio in Wai‘anae

Ahupua‘a, Wai‘anae District, which supplied Kualoa residents with

most of their volcanic glass. In addition, Kualoa residents had access to

volcanic glass from the Pu‘uwa‘awa‘a source on Hawai‘i Island, some

300 km away.

The results of the paired technological and geochemical sourcing

analyses are interpreted as supporting the laissez-faire view of political

development. Nevertheless, questions remain and it is recommended

that similar technological and geochemical sourcing analyses be carried

out at locations with well-dated contexts capable of tracking changes

over time in the chaîne opératoire for volcanic glass.

2. Volcanic glass studies in Hawai‘i

Volcanic glass artifacts were

ﬁrst

recognized and reported by Bishop

Museum archaeologists in the early 1960’s(Soehren, 1962). Since then,

archaeologists have carried out technological, edge-damage, dating,

and sourcing studies (see Supplementary Materials).

Technological and edge-damage investigations culminated in the

early 1980’s with an elaborate study of 675 ﬂakes and 25 cores of

volcanic glass collected from seven sites on the leeward side of Hawai‘i

Island (Schousboe et al., 1983, 353). Qualitative and quantitative ob-

servations on 23 attributes were recorded with the aid of a 20× bi-

nocular microscope. Observations focused on details of ﬂaking plat-

forms, characteristics of scars on the dorsal face of ﬂakes, and edge

damage. The study concluded that volcanic glass was ﬂaked using

“hard-hammer percussion” (Schousboe et al., 1983, 362) in a reduction

sequence characterized as a “continuum where initially hand-held cores

were eventually reduced to a point where a bipolar technique became

necessary for further ﬂake removal” (Schousboe et al., 1983, 362).

Cores were turned frequently during reduction, presumably to identify

or set up suitable ﬂaking platforms (Schousboe et al., 1983, 363). The

detailed study of edge damage left the authors unable “to make any

deﬁnitive statements on the functions of the … volcanic glass artifacts”

(Schousboe et al., 1983, 368). The high cost/beneﬁt ratio indicated by

this result frustrated the authors’“hope that other archaeologists in

Hawai‘i will take the time to examine their volcanic-glass assemblages

with this perspective” (Schousboe et al., 1983, 368).

Technological studies have contributed to sourcing projects that

interpret the distribution of material away from the source in terms of

trade and exchange. In these studies, direct access is indicated by the

presence of cortex on ﬂakes and by large ﬂakes, whereas movement of

used cores in down-the-line exchange is indicated by rare cortex and

small ﬂakes (Putzi et al., 2015; McCoy et al., 2011).

Hydration rind dating of volcanic glass was introduced to Hawaiian

archaeology in the early 1970’s(Morgenstein and Rosendahl, 1976;

Morgenstein and Riley, 1974; Barrera and Kirch, 1973

), but the tech-

nique

was criticized (Graves and Ladefoged, 1991; Olson, 1983) and

most Hawaiian archaeologists now regard it as “scientiﬁc noise”

(Tuggle, 2010, 177). A recent attempt to resurrect the hydration rind

dating of Pu‘uwa‘awa‘a volcanic glass yielded dates that defy archae-

ological interpretation (Stevenson and Mills, 2013).

By the late 1970’s, archaeologists working on Hawai‘i Island had

learned that hand samples of Pu‘uwa‘awa‘a volcanic glass were iden-

tiﬁable in hand sample (Schousboe et al., 1983; Reeve, 1983; Kirch,

1979). Early attempts to distinguish volcanic glass sources chemically

(Weisler, 1990; Olson, 1983) were followed by establishment of an

EDXRF (Energy-dispersive X-ray ﬂuorescence) facility at the University

of Hawai‘i at Hilo, which used the technique to distinguish Pu‘u-

wa‘awa‘a from unknown sources chemically consistent with Mauna Loa

and Kīlauea lava ﬂows on Hawai‘i Island (Lundblad et al., 2013).

Fig. 1. The Hawaiian Islands with traditional district boundaries and ahupua‘a names mentioned in the text.

N.J. DiVito, et al.

Journal of Archaeological Science: Reports 30 (2020) 102117

EDXRF was used to establish that the proportion of Pu‘uwa‘awa‘a vol-

canic glass in an archaeological collection declines with distance from

the source (McCoy et al., 2011). Later, a distance decay function was ﬁt

to the Hawai‘i Island data (Putzi et al., 2015).

3. Archaeological investigations at Kualoa

Kualoa is located at the northern end of Kāne‘ohe Bay along the

windward coast of O‘ahu Island. It is the northernmost ahupua‘a,or

land, in the Ko‘olaupoko district. Kualoa is a small ahupua‘a consisting

of a narrow coastal plain and a sandy peninsula, with steep cliﬀs tow-

ering above the landscape (Fig. 2). Nevertheless, it is famous in Ha-

waiian tradition as a sacred place, a symbol of island sovereignty, and

an established place of refuge.

The sandy peninsula of Kualoa where the volcanic glass pieces were

collected is today a public park owned and managed by the City and

County of Honolulu. In the mid-nineteenth century there were two

settlements in Kualoa, one near the beach on the north and another on

the alluvium adjoining the open water of Mōli‘i Pond. Sugarcane was

cultivated during the 1860’s. After the sugarcane operation folded, the

land was used to pasture horses and cattle. When the ﬁrst archae-

ological survey of Kualoa was completed as part of an island-wide

survey no trace could be found of two temples identiﬁed by Hawaiian

tradition (McAllister, 1933). The volcanic glass analyzed here was

collected by City and County archaeologists in the 1970’s and also

during University of Hawai‘i archaeological ﬁeld schools in 1977, 1983,

and 1984 (Gunness, 1993), primarily at seven locations that roughly

coincide with the locations of the mid-nineteenth century settlements

(see Fig. 2). The archaeological collections were augmented by a small

geological collection from dikes exposed on the east side of Mokoli‘i

Island by Robert D. Connolly.

4. Materials and methods

Paired technological and geochemical sourcing analyses were car-

ried out on 1258 pieces of volcanic glass on loan from the City and

County of Honolulu.

Technological observations included: length, width, and thickness

measured to the nearest 0.01 mm with a digital caliper; weight mea-

sured to the nearest 0.1 g with a digital scale; formal classiﬁcation of

the volcanic glass piece as an unworked nodule, core, ﬂake, edge-

altered ﬂ

ake, or debris; number of facets; presence/absence of macro-

scopic edge alteration; completeness; and identiﬁcation of primary,

secondary, and tertiary cortex; along with notes about the condition of

the piece and its possible interpretation (see Supplementary Materials).

Non-destructive geochemical analysis was carried out with a

ThermoNoran QuantX

EDXRF spectrometer at the Geoarchaeology

Laboratory of the University of Hawaii at Hilo. Data were acquired for

18 elements: Na, Mg, Al, Si, K, Ca, Ti, V, Mn, Fe, Ni, Cu, Zn, Rb, Sr, Y,

Zr, and Nb. The elements Rb, Sr, Y, Zr, and Nb most often exhibit the

best analytical precision with EDXRF (Mills et al., 2010). A scatterplot

of the concentration of elements Sr and Zr was used to identify clusters

and distinguish 11 compositional groups by eye. A principal compo-

nents analysis yields results that support the grouping (see

Supplementary Materials). The low cost and high throughput of EDXRF

analysis have made it a valuable method for revealing general patterns

in geochemistry within a volcanic glass collection. The results of EDXRF

analysis can be used to guide application of expensive, destructive,

and/or time-consuming techniques (Kirch et al., 2012).

Twenty-four broken and/or unworked pieces of volcanic glass were

subsequently analyzed at the School of Ocean and Earth Science and

Technology (SOEST) Microprobe Facility at the University of Hawai‘iat

Mānoa. Data were acquired for SiO

, TiO

,Al

, FeO*, MnO, MgO,

CaO, Na

O, K

O, and P

. These data support and augment the EDXRF

analysis. They were compared with published analyses to identify

volcanic glass sources (see Supplementary Materials).

5. Results

The 1258 volcanic glass pieces in the technological analysis in-

cluded 58 unworked nodules, 3 cores, 609 ﬂakes with macroscopic

evidence for edge alteration, 522 ﬂakes without macroscopic evidence

for edge alteration, and 69 pieces of debris. Two locations yielded the

bulk of this material. The University of Hawai‘i ﬁeld school excavations

in area 2B-1 yielded 744 pieces of volcanic glass, and the surface col-

lection from Area 1B contributed 439 pieces. In contrast, the other lo-

cations each yielded between 2 and 31 pieces of volcanic glass. The

distribution of classes across the collection areas is relatively even.

Nevertheless, nodules and debris are more common in the excavated

collection from area 2B-1 than in the surface collection from area 1B,

and all three cores came from the excavations in area 2B-1.

The size of a volcanic glass piece varies according to class. Cores are

Fig. 2. Collections locations at Kualoa, with places mentioned in the text.

N.J. DiVito, et al.

Journal of Archaeological Science: Reports 30 (2020) 102117

on average slightly larger than the unworked nodules. The mean

complete ﬂake with macroscopic evidence for edge alteration is longer,

wider, and thicker than the mean complete ﬂake with an unaltered edge

(Fig. 3). Flakes without evidence for edge alteration are close in size to

pieces of debris.

The size of complete edge-altered ﬂakes varies according to whether

the ﬂake exhibits primary, secondary, or tertiary cortex (Fig. 4). The

length and width of ﬂakes with primary and secondary cortex cannot be

distinguished, but ﬂakes with tertiary cortex are shorter and narrower.

In contrast, thickness declines through the sequence.

A second dimension of variation is evident in the size relationship

between complete edge-altered ﬂakes with tertiary cortex and complete

unaltered ﬂakes with secondary cortex. Edge-altered ﬂakes with ter-

tiary cortex are longer and wider than unaltered ﬂakes with secondary

cortex, but thickness can’t be distinguished.

A third dimension of variation compares the numbers of altered and

unaltered ﬂakes at each stage. Flakes with primary and secondary

cortex more frequently show edge alteration than not. In contrast,

ﬂakes with tertiary cortex more frequently have unaltered edges (see

Supplementary Materials).

Flakes were typically discarded before they broke. 894 ﬂakes in the

collection are complete and 235 ﬂakes are broken. Breakage patterns

vary according to edge alteration; ﬂakes without macroscopic evidence

for edge alteration are more likely broken than ﬂakes with macroscopic

evidence for edge alteration (see Supplementary Materials).

Geochemical source analysis indicates that some of the vitreous

material identiﬁed by archaeologists as volcanic glass is either not glass

or is not completely vitriﬁed (see Supplementary Materials). EDXRF

analysis identiﬁed 44 pieces whose chemistry is outside the known

range of Hawaiian rocks; these pieces likely include bits of burned

bone, burned shell, charcoal, and possibly coal. Microprobe analysis

identiﬁed one piece—the proximal end of a ﬂake with tertiary cor-

tex—as metallic, most likely the Fe hydroxide goethite, a weathering

product of Hawaiian volcanic rocks. In addition, 2 of the 24 pieces

submitted for microprobe analysis are partially devitriﬁed, showing

evidence for incipient crystallization (Fig. 5). One of the partially de-

vitriﬁed pieces is the proximal end of a ﬂake with macroscopic evidence

of edge alteration, and the other was a nodule that appeared to have

been used as a scraper but had not had ﬂakes removed from it.

The 1258 volcanic glass pieces in the geochemical source analysis

were divided into 10 groups by the EDXRF analysis. The collection is

dominated by Groups 9 and 4, which contain 638 and 535 pieces, re-

spectively. 36 pieces were assigned to source Group 8. Source groups 2,

3, 5–7, 10, and 11 are all small, ranging from 1 to 12 pieces (see

Supplementary Materials).

The EDXRF groups are distributed evenly among the collection

areas at Kualoa. The two large collections from areas 2B-1 and 1B each

contain volcanic glass pieces from source groups 3–11. EDXRF Group 9

is more numerous than Group 4 in each of these large collections,

consistent with its representation in the collection as a whole (see

Supplementary Materials).

EDXRF Groups 9 and 4 are each dominated by ﬂakes with and

without macroscopic edge damage. Group 4 has proportionally more

debris than Group 9. There are unworked nodules assigned to 9 of the

groups; only the sparsely populated Group 5 does not include unworked

nodules of volcanic glass (see Supplementary Materials).

The complete edge-altered ﬂakes assigned to Groups 4 and 9 diﬀer

little in size (Fig. 6).

A speciﬁc source has been identiﬁed for 2 of the 11 groups dis-

tinguished by the EDXRF analysis (see Supplementary Materials). Mi-

croprobe analysis conﬁrmed that a piece of volcanic glass from Group

11 derives from the distinctive Pu‘uwa‘awa‘a source on Hawai‘i Island,

a result that lends further support to the claim that EDXRF analysis

conﬁdently identiﬁes materials from this distinctive source.

Microprobe analysis indicates 11 volcanic glass pieces assigned to

Group 9 as deriving from a single silicic glass source at Pu‘uKa‘īlio in

the Wai‘anae Range. One volcanic glass piece assigned to Group 9 was

identiﬁed as deriving from a basaltic glass source in the Wai‘anae

Range, as were one piece assigned to Group 6 and another piece as-

signed to Group 4 (see Supplementary Materials).

Volcanic glass from the dikes on Mokoli‘i Island would be assigned

to Group 4. However, microprobe analyses of glasses in Group 4 are

suﬃciently diﬀerent from one another to suggest at least 7 separate

volcanic glass sources, most likely in the Ko‘olau Range, and at least one

source in the Wai‘anae Range. Basaltic glass sources on other islands

might also have contributed material to the Kualoa collection. In ad-

dition, the microprobe analysis assigns EDXRF Group 3 to a basaltic

source in the Ko‘olau

Range.

6. Discussion

Interpretation of the results of the paired technological and geo-

chemical sourcing analyses of 1258 pieces of volcanic glass from Kualoa

attempts to reconstitute the chaîne opératoire for an expedient tool that

is nearly ubiquitous in Hawaiian archaeological collections.

Technological analyses are best interpreted as a reduction sequence

involving small nodules of volcanic glass ﬂaked with a hard hammer

using both hand-held and bipolar techniques, in which the core was

turned frequently to ﬁnd a suitable striking platform (Schousboe et al.,

1983, 362). Flake removal typically yielded some debris along with

ﬂakes of various sizes. Macroscopic observation of edge damage was

used in the technological analysis to distinguish between ﬂakes selected

Fig. 3. Box and whisker plots comparing the sizes of 609 complete volcanic glass ﬂakes with macroscopic evidence for edge alteration, 522 complete ﬂakes without

macroscopic evidence for edge alteration, and 69 pieces of debris. Box widths are proportional to the number of instances. The notches in the boxes approximate 95%

conﬁdence intervals for comparing medians.

N.J. DiVito, et al.

Journal of Archaeological Science: Reports 30 (2020) 102117

for use and those that were used for tasks that didn’t result in macro-

scopic traces of edge damage or were discarded without being used.

This interpretation is supported by the marked diﬀerences in size be-

tween the edge-altered and unaltered complete ﬂakes, where a strong

bias for large ﬂakes is evident.

Nevertheless, this interpretation is potentially tempered by the

possibility that edges might be damaged by some activity other than

use. In particular, trampling of discarded ﬂakes might produce edge

damage similar to that produced when a ﬂake is used for cutting or

scraping tasks. The trampling problem has been studied experimentally

by archaeologists who have identiﬁed three variables that inﬂuence the

probability an edge will be damaged by trampling:

(i) stiﬀness of substrate on which the ﬂakes are deposited

(ii) density of ﬂakes in the deposit, and

(iii) ﬂake size (Giﬀord-Gonzalez et al., 1985; McBrearty et al., 1998).

Experiments show consistently that ﬂakes are more often damaged on

stiﬀ substrates such as loam, where trampling does not cause ﬂakes to

penetrate the substrate, than on softer substrates such as un-

consolidated sand, where trampling pushes ﬂakes into the substrate

(Giﬀord-Gonzalez et al., 1985; McBrearty et al., 1998). The volcanic

glass collections from Kualoa were all recovered from deposits of

Fig. 4. Box and whisker plots comparing the sizes of 352 complete volcanic glass ﬂakes with primary cortex, 225 complete volcanic glass ﬂakes with secondary

cortex, and 319 complete volcanic glass ﬂakes with tertiary cortex. Box widths are proportional to the number of instances. The notches in the boxes approximate

95% conﬁdence intervals for comparing medians.

Fig. 5. Backscatter electron image showing partially devitriﬁed volcanic glass

with evidence for incipient crystallization; note scale bar in lower right hand

corner. Sample 443–221–1 is the proximal end of a broken ﬂake with macro-

scopic evidence for edge alteration.

N.J. DiVito, et al.

Journal of Archaeological Science: Reports 30 (2020) 102117

unconsolidated calcareous sand, a soft substrate. The experiments also

show that most of the damage caused by trampling occurs when two

ﬂakes come into contact with one another or when a ﬂake comes into

contact with another hard object. Thus, edge damage from trampling is

relatively common at chipping stations where there is a dense dis-

tribution of ﬂakes and debris and less common in situations where

ﬂakes are sparsely distributed. In one experiment, densities ranged from

100 to 500 artifacts per m

(McBrearty et al., 1998, 112), while another

experiment used a ﬁxed density of 250 artifacts per m

(Giﬀord-

Gonzalez et al., 1985, 805). At Kualoa, densities of volcanic glass ar-

tifacts are diﬃcult to establish because collections in any one area were

typically made over long periods of time and in circumstances ranging

from surface collection after grading and grubbing to controlled ex-

cavation by ﬁeld school students. In any case, volcanic glass density

ranged from about 0.2 per m

in Area IB, which was graded and

grubbed and subsequently subject to limited test excavations, to about

1–2 per m

in areas 2B, 2C, and 2D where construction-related dis-

turbance was augmented with large scale excavations carried out by

university ﬁeld schools. In all cases, volcanic glass densities at Kualoa

were orders of magnitude lower than the experimental situations. Fi-

nally, the experiments demonstrated that ﬂakes longer than 2 cm were

fractured by trampling much more frequently than shorter ﬂakes

(Giﬀord-Gonzalez et al., 1985, 807), especially on a soft substrate. Edge

damage to the shorter ﬂakes was extremely rare on both stiﬀ and soft

substrates (Giﬀord-Gonzalez et al., 1985, 814). The vast majority of

volcanic glass artifacts from Kualoa are small, with only a few ﬂakes

longer than 2 cm (see Fig. 3). The diminutive size of these artifacts

makes them unlikely candidates for damage from trampling.

By all measures—substrate stiﬀness, ﬂake density, and ﬂake si-

ze—the volcanic glass artifacts from Kualoa are unlikely to show da-

mage from trampling that might be confused with edge damage caused

by use wear. Thus, comparisons of the sizes of ﬂakes with and without

edge damage in the Kualoa assemblage can be interpreted as a result of

habitual practices of Hawaiian craftsmen who selected some ﬂakes for

uses that left macroscopic damage to the cutting edge, such as working

wood or bone, and either discarded others or used them for working

softer materials in such a way that macroscopic edge damage did not

result (Young and Bamforth, 1990; Shen, 1999).

The selection habits of volcanic glass producers and users can be

tracked through the reduction sequence by comparing the sizes of edge-

altered and unaltered ﬂakes with primary, secondary, and tertiary

cortex (see Fig. 4). Selection for length and width appear to have op-

erated more strongly than selection for thickness, as indicated by the

fact that edge-altered tertiary ﬂakes are longer and wider than un-

altered secondary ﬂakes. Flakes longer than 8 mm or wider than 7 mm

were typically selected for use, while ﬂakes shorter or narrower than

this were not selected. This pattern of selection might be expected for

an expedient tool industry, where a ﬂake had to be large enough to grip

securely between the thumb and foreﬁnger. In this interpretation, ﬂakes

smaller than 8 × 7 mm were too small to grip securely for cutting and

scraping tasks. If this is correct, then a core of volcanic glass would have

been considered exhausted and ready for discard when it was no longer

able to yield a ﬂake of this size.

The selection habits of volcanic glass producers and users also can

be tracked through the reduction sequence by comparing the numbers

of edge-altered and unaltered ﬂakes with primary, secondary, and ter-

tiary cortex (see Fig. 4). The proportion of

ﬂakes

selected for use de-

creases through the reduction sequence. About 2 of every 3 ﬂakes with

primary cortex are selected for tasks that leave macroscopic evidence of

edge damage. The selection rate drops to 3 out of 5 among ﬂakes with

secondary cortex. By the time the cortex is removed, fewer than half the

ﬂakes show evidence of edge damage. The selection habits of volcanic

glass workers clearly illustrate the process by which a core was ex-

hausted.

Another potential reason for selection of larger ﬂakes for tasks that

result in macroscopic edge damage appears to be the tendency of

smaller ﬂakes to break during use. The implication here is that volcanic

glass ﬂakes used as tools often broke relatively quickly, before macro-

scopic edge damage occurred.

EDXRF and microprobe analyses conﬁrm the Pu‘uwa‘awa‘a origin of

3 pieces of volcanic glass, one of which was identiﬁed in hand sample

and reported as “the ﬁrst archaeologically documented evidence of

inter-island transport of what might be called a ‘high-status’ consumer

good” (Gunness, 1993, 58). The 3 pieces identiﬁed in the analysis is one

fewer than the 4 pieces predicted by the distance decay function

(Fig. 7). The predictive accuracy of the distance decay function extends

Fig. 6. Box and whisker plots comparing the sizes of 210 complete edge-altered ﬂakes assigned to Group 4 and 269 complete edge-altered ﬂakes assigned to Group 9.

Box widths are proportional to the number of instances. The notches in the boxes approximate 95% conﬁdence intervals for comparing medians.

Fig. 7. Distribution of Pu‘uwa‘awa‘ a volcanic glass away from the source. The

distance decay function is deﬁned as

=−

exp d(5.104315 0.035707 * )

, where p

is the percentage of Pu‘uwa‘awa‘a volcanic glass in the assemblage, and dis-

tance, d, is measured as the shortest path in km from Pu‘uwa‘awa‘a to the ﬁnd

spot. Sources: McCoy et al. (2011),Putzi et al. (2015).

N.J. DiVito, et al.

Journal of Archaeological Science: Reports 30 (2020) 102117

its utility to instances of inter-island movement of volcanic glass across

political boundaries. The Hawai‘i Island data showed that volcanic glass

moved across ahupua‘a and moku district boundaries, either by trans-

port over land or by canoe (Putzi et al., 2015). Identiﬁcation of 3 pieces

of Pu‘uwa‘awa‘a glass in the Kualoa collection show that volcanic glass

also moved across mokupuni island boundaries by canoe. The unworked

nodule of Pu‘uwa‘awa‘a glass recovered at Kualoa suggests that vol-

canic glass workers at Kualoa enjoyed direct access to the Pu‘uwa‘awa‘a

source, and were not dependent on down-the-line exchange for access

to this material.

EDXRF and microprobe analyses support the inference that most of

the volcanic glass in the Kualoa collection derives from a source in the

vicinity of Pu‘uKa‘īlio in Wai‘anae Ahupua‘a, a distance of about 32 km

from Kualoa. Somewhat surprisingly, the distance decay function for

Pu‘uwa‘awa‘a correctly predicts the proportion of the Pu‘uKa‘īlio

source in the Kualoa collection. The distance decay function predicts

that a source 32 km distant should make up 52 percent of the volcanic

glass collection; the observed proportion of Group 9 glass at Kualoa is

51 percent. Assignment to Group 9 of 29 unworked volcanic glass no-

dules indicates Kualoa volcanic glass workers enjoyed direct access to

the Pu‘uKa‘īlio source, too, and did not have to rely on down-the-line

exchange.

These results suggest the distance decay function might be applied

to the distribution of volcanic glass from sources other than

Pu‘uwa‘awa‘a. If so, archaeologists might claim evidence for a wide-

spread tradition of social relations unaﬀected by development of state

political institutions (Renfrew, 1977). Because the Pu‘uKa‘īlio source

appears to be reliably identiﬁ

ed by EDXRF, it should be possible to test

the distance decay function with other large volcanic glass assemblages.

7. Conclusions

The results of paired technological and geochemical sourcing ana-

lyses of 1258 volcanic glass pieces from Kualoa are interpreted to in-

dicate that the quotidian habits of Kualoa volcanic glass workers con-

nected them directly with resources well outside the ahupua‘a

community. Development of a political hierarchy in the islands does not

appear to have inﬂuenced these habits. If this characterization is cor-

rect, then Hawai‘i conforms to the typical situation in pre-capitalist

social formations in which leaders concern themselves with controlling

the ways in which status is acquired and how wealth assets are dis-

tributed, rather than the quotidian habits of common folk and the

distribution of consumables, such as volcanic glass (Rowlands, 1982;

Goldman, 1970). The picture that emerges from the paired technolo-

gical and geochemical sourcing analyses of the Kualoa volcanic glass

collection is of a society in which people traveled widely and enjoyed

support for their productive activities with unfettered access to sources

of volcanic glass.

This interpretation is tempered somewhat by the tradition of Kualoa

as a sacred place. A single paired technological and geochemical

sourcing analysis cannot establish the pattern of distribution away from

the Pu‘uKa‘īlio source, and subsequent inquiry at other locations might

show Kualoa to be a special case where the high status of ahupua‘a

residents entailed preferred access to resources such as volcanic glass.

Also, the Kualoa volcanic glass derives from poorly dated contexts that

frustrate the possibility of tracking changes in the chaîne opératoire for

volcanic glass. These questions about the Kualoa collection can be ad-

dressed by paired technological and geological sourcing analyses at

other Hawaiian sites with large volcanic glass collections from well-

dated contexts.

8. Funding sources

This work was partially supported by the National Science

Foundation [BCS-1427950] and by T.S. Dye & Colleagues,

Archaeologists.

Acknowledgments

We thank the following people, without whom the work reported

here would not have been completed: Michele Nekota, Director,

Department of Parks and Recreation, City and County of Honolulu;

Linda K. Liu, Department of Parks and Recreation; Teo Clemens and

Will Gardner, Kualoa Regional Park; Caleb Fechner, Paciﬁc Legacy; and

Rona Ikehara-Quebral and Alex Morrison, International Archaeological

Research Institute.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the

online version, at https://doi.org/10.1016/j.jasrep.2019.102117.

References

Athens, J.S., Rieth, T.M., Dye, T.S., 2014. A paleoenvironmental and archaeological

model-based age estimate for the colonization of Hawai‘i. Am. Antiq. 79 (1),

144–155.

Barrera Jr., W., Kirch, P.V., 1973. Basaltic glass artefacts from Hawaii: their dating and

prehistoric uses. J. Polynesian Soc. 82, 176–187.

Giﬀord-Gonzalez, D.P., Damrosch, D.B., Damrosch, D.R., Pryor, J., Thunen, R.L., 1985.

The third dimension in site structure: an experiment in trampling and vertical dis-

persal. Am. Antiq. 50 (4), 803–818.

Goldman, I., 1970. Ancient Polynesian Society. University of Chicago Press, Chicago.

Graves, M.W., Ladefoged, T.N., 1991. The disparity between radiocarbon and volcanic

glass dates: new evidence from the island of Lāna‘i, Hawai‘i. Archaeol. Oceania 26,

70–77.

Gunness, J.L., 1993. The Kualoa archaeological research project, 1975 –1985: a brief

overview. Hawaiian Archaeol. 2, 50–71.

Hommon, R.J., 2013. The Ancient Hawaiian State: Origins of a Political Society. Oxford

University Press, Oxford.

Kirch, P.V., 1979. Marine Exploitation in Prehistoric Hawai‘i: Archaeological

Investigations at Kalāhuipua‘a, Hawai‘i Island. No. 29 in Paciﬁc Anthropological

Records. Anthropology Department, B.P. Bishop Museum, Honolulu.

Kirch, P.V., 2010. How Chiefs Became Kings: Divine Kingship and the Rise of Archaic

States in Ancient Hawai‘i. University of California Press, Berkeley, CA.

Kirch, P.V., Mills, P.R., Lundblad, S.P., Sinton, J., Kahn, J.G., 2012. Interpolity exchange

of basalt tools facilitated via elite control in Hawaiian archaic states. Proc. Nat. Acad.

Sci. 109 (4), 1056–1061.

Lundblad, S.P., Mills, P.R., McCoy, M.D., Kahn, J.G., Mulliken, K., Kaylor, D., 2013.

Identiﬁcation of volcanic glass sources inferred from geochemical analysis of artifacts

on leeward Hawai‘i Island. In: Summerhayes, G.R., Buckley, H. (Eds.), Paciﬁc

Archaeology: Documenting the Past 50,000 Years. Papers from the 2011 Lapita

Paciﬁc Archaeology Conference. No. 25 in University of Otago Studies in

Archaeology. Dunedin, NZ, pp. 67–75.

McAllister, J.G., 1933. Archaeology of Oahu. No. 104 in B.P. Bishop Museum Bulletin.

Bishop Museum Press, Honolulu.

McBrearty, S., Bishop, L., Plummer, T., Dewar, R., Conard, N., 1998. Tools underfoot:

human trampling as an agent of lithic artifact edge modiﬁcation. Am. Antiq. 63 (1),

108–129.

McCoy, M.D., Mills, P.R., Lundblad, S.P., Rieth, T.M., Kahn, J.G., Gard, R., 2011. A cost

surface model of volcanic glass quarrying and exchange in Hawai‘i. J. Archaeol. Sci.

38, 2547–

2560.

Mills, P.R., Lundblad, S.P., Field, J.S., Carpenter, A.B., McElroy, W., Ross, P., 2010.

Geochemical sourcing of basalt artifacts from Kaua‘i, Hawaiian Islands. J. Archaeol.

Sci. 37, 3385–3393.

Morgenstein, M., Riley, T.J., 1974. Hydration-rind dating of basaltic glass: a new method

for archaeological chronologies. Asian Perspect. 17, 145–159.

Morgenstein, M., Rosendahl, P., 1976. Basaltic glass hydration dating in Hawai‘i. In:

Taylor, R.E. (Ed.), Advances in Obsidian Glass Studies. Noyes Press, Park Ridge, NJ,

pp. 141–164.

Olson, L., Feb. 1983. Hawaiian volcanic glass applied “dating” and “sourcing”:

Archaeological context. In: Clark and Kirch, pp. 325–340.

Putzi, J.L., DiVito, N.J., Sholin, C.E., Mills, P.R., Lundblad, S., Camara, B., Dye, T.S., 2015.

Alternative models of volcanic glass quarrying and exchange in Hawai‘i. J.

Archaeolog. Sci.: Rep. 2, 341–352.

Reeve, R.B., Feb. 1983. Archaeological investigations in Section 3. In: Clark, J.T., Kirch, P.

V. (Eds.), Archaeological Investigations of the Mudlane-Waimea-Kawaihae Road

Corridor, Island of Hawai‘i: An Interdisciplinary Study of an Environmental Transect.

No. 83–1 in Departmental Report Series. Anthropology Department, B.P. Bishop

Museum, Honolulu, pp. 181–239.

Renfrew, C., 1977. Alternative models for exchange and spatial distribution. In: Exchange

Systems in Prehistory. Studies in archeology. Academic Press, New York, pp. 71–90.

Rowlands, M.J., 1982. Processual archaeology as historic social science. In: Renfrew, C.,

Rowlands, M.J., Segraves, B.A. (Eds.), Theory and Explanation in Archaeology: The

Southampton Conference. Academic Press, New York, NY, pp. 155–174.

Schousboe, R., Riford, M., Kirch, P.V., Feb. 1983. Volcanic glass ﬂaked stone artifacts. In:

Clark, J.T., Kirch, P.V. (Eds.), Archaeological Investigations of the Mudlane-Waimea-

N.J. DiVito, et al.

Journal of Archaeological Science: Reports 30 (2020) 102117

Kawaihae Road Corridor, Island of Hawai‘i: An Interdisciplinary Study of an

Environmental Transect. No. 83–1 in Departmental Report Series. Anthropology

Department, B.P. Bishop Museum, Honolulu, pp. 348–370.

Shen, C., 1999. Were “utilized ﬂakes” utilized? an issue of lithic classiﬁcation in Ontario

archaeology. Ontario Archaeol. 68, 63–73.

Soehren, L., 1962. Archaeological excavations at City of Refuge National Historical Park,

Honaunau, Kona, Hawaii. Prepared for United States National Park Service.

Anthropology Department, B.P. Bishop Museum, Honolulu.

Stevenson, C.M., Mills, P., 2013. A chronometric tool for Hawaiian archaeology: the

hydration dating of Pu‘uwa‘awa‘a trachytic glass. J. Archaeol. Sci. 40, 405–415.

Tuggle, H.D., 2010. Lady Mondegreen’s hopes and dreams: three brief essays on inference

in Hawaiian archaeology. In: Dye, T.S. (Ed.), Research Designs for Hawaiian

Archaeology. No. 3 in Special Publications. Society for Hawaiian Archaeology,

Honolulu, Ch. 3, pp. 157–184.

Weisler, M.I., 1990. Sources and sourcing of volcanic glass in Hawai‘i: implications for

exchange studies. Archaeol. Oceania 25 (1), 16–23.

Young, D., Bamforth, D.D., 1990. On the macroscopic identiﬁcation of used ﬂakes. Am.

Antiq. 55 (2), 403–409.

N.J. DiVito, et al.

Journal of Archaeological Science: Reports 30 (2020) 102117