14
BIRD BONES FROM THE TO'AGA SITE:
PREHISTORIC LOSS OF SEABIRDS AND MEGAPODES
DAVID W. STEADMAN
INTRODUCTION
A      S PART OF A LONG-TERM program to reconstrUct
the natural distribution and diversity of bird-
life in the South Pacific (figure 14.1), I have sought
bird bones from archaeological sites in many
different archipelagos. Until recently, the islands of
Samoa have not been represented in this data base.
In 1987, T. L. Hunt and P. V. Kirch conducted
excavations at the To'aga site (AS-13-1), Ofu Island,
American Samoa yielding a small sample of bird
bones representng at least six taxa (Steadman 1990).
This first glimpse of the prehistoric avifauna of
Samoa showed that at least two species (a shearwa-
ter and petrel) had been lost on Ofu since the first
anrival of humans more than 3000 years ago. Hunt
and Kirch expanded the excavations at To'aga in
1989. This paper reports the entire bird bone
assemblage from both the 1987 and 1989 field
seasons. The sample of bird bones from To'aga now
consists of at least fifteen taxa (table 14.1) and
provides a more thorough, although still far from
complete, portrayal of the birdlife of ancient Samoa.
The bird bones from To'aga have been cata-
logued in the University of Washington Burke
Museum (UWBM) Fossil Bird Collection. Com-
parative skeletal or oological specimens are from
UWBM, the British Museum (Natural History)
(BMNH), the New York State Museum (NYSM),
and the United States National Museum of Natural
History, Smithsonian Institution (USNM). hi the
species accounts that follow, "Unit" refers to meter
square excavations (designated with Arabic numer-
als). Roman numerals refer to stratigraphic layers.
For details of the stratigraphy, chronology, material
culture, and non-bird faunal assemblages of the
To'aga site, see Hunt and Kirch (1988), Kiivh et al.
(1989, 1990), and various chapters in this volume.
Unless stated otherwise, the modem distribu-
dons of species within American Samoa are taken
from the excellent surveys conducted in 1975-1976
by Amerson et al. (1982ab) and in 1986 by
Engbring and Ramsey (1989). Modem distributions
for elsewhere in Polynesia are taken from Pratt et al.
(1987). The prehistoric records are from Steadman
(1989a) and from unpublished data from my recent
research; islands preceded by an asterisk (*) repre-
sent extirpated populations.
SPECIES ACCOUNTS
Order Procellariiformes
Family Procellanidae
Pwffinuspacificus (Wedge-tailed Shearwater)
MATERIAL. Humerus (UWBM 1680), T9/
500E, Unit 21, fiB. Five ulnae (UWBM 1244, 1251,
218    The To'aga Site
Table 14.1
Bird Bones from the To'aga Site
Number of Identified Bones
Taxa
Seabirds
*Puffinus pacificus
*Puffnus iherminieri
*Puffinus griseus
*Pterodroma rostrata
*Pterodroma sp., size of P. heraldica
(*)Procellariidae sp.
*Sula sula
Fregata sp.
Anous stolidus
Gygis alba
Steminae sp.
Landbirds
Egretta sacra
Numenius tahitiensis (M)
Gallus gallus (I)
*Megapodius sp.
Gallicolumba stairii
TOTAL
Total Seabirds (species/bones)
Total Landbirds (species/bones)
Total Native Landbirds, without IM (species/bones)
% of Bones from Extirpated Species (without IM)
11
2
15
6
2
9
1
2
1
1
1
1
1
16
2
3
15n4
10/51
5/23
4/7
85%
I=Introduced species
M=Migrant Species
* Extirpated on Ofu
(*) represents extirpated species, but not necessarily different from those already listed.
1256, 1678, 1679), Unit 9, IIB; Unit 5, IIC; Unit 14,
llIa,4; 19/500E, Unit 21, IIB (2 bones). Radius
(UWBM 1664), T9/500E, Unit 20, IIIB.
Carpometacarpus (UWBM 1630), T3I200W. Unit
27, fI1A. Tibiotarsus (UWBM 1642), T5/100E, Unit
28, HIB. Two pedal phalanges (UWBM 1246, 1248),
Unit 4, IIB; Unit 1, IIB.
REMARKS. This tropical shearwater is rarely
noted at sea today in American Samoa It may nest
on Pola Islet (off Tutuila) and on Ta'u, but this has
not been confirmed. There are no previous rcords
from Ofu. It is widespread in Polynesia today,
although nesting islands are few. Other Polynesian
archaeological records of PwJnuspac~lcus are
from *Ua Huka and *Tahuata (Marquesas),
*HuAhine (Society Islands), *Lifuka and * Eua
(Tonga), and Tikopia and Anuta (Solomon Islands).
Puojinus lherminieri (Audubon's Shearwater)
MATERIAL. 2 ulnae (UWBM 1651, 1671),
T9/500E, Unit 23, IIB; T9/500E, Unit 21, IIB.
REMARKS. Within American Samoa, Puffinus
iherminueri nests only on Tau, where at least 200
birds breed in the cloud forest, and on Tutila
(colony size unknown). It is uncommon at sea and
has not been recorded previously on Ofu. Like P.
pacificus, P. lherminieri nests today on relatively
few islands scattered across Polynesia. Other
Polynesian archaeological records of P. Iherminieri
are from *Nuku Hiva, Ua Pou, *Ua Huka, *Hiva Ok,
and *Tahuata (Marquesas), *Huahine (Society
Prehistoric Loss of Seabirds and Megapodes  219
I  I
N
o
0C C
C,5
J .
4. *0
0
m
0~~~
z~~~~~~~ o
w   -J  ~~~~~~~~~~-J
*.   *  X
0  C)(
00J
(4
t     ze C  s*  * 4
0  -0
.   a Z
z~~~~~
I-.a  ~~~~~~~~ .00
: Y  0 - J
u0
0  Z ;'=   | *oa
uot~~~~~~~~~~; J   i
Z
'0  w   0I
a -J
4    - 4    -
z~~~~~
0.
-  z
z
w
Z
9~~~~~~~~~c
*r.
220     The Tolaga Site
Islands), *Mangaia (Cook Islands), Eua (Tonga),
and Tikopia and *Anuta (Solomon Islands).
Puffinus griseus (Sooty Shearwater)
MATERIAL. Sternum (UWBM 1681), T9/
500E, Unit 21, IIB. Coracoid (UWBM 1645), T5/
100E, Unit 29, RIB. Scapula (UWBM 1245), Unit
9, IIB. Three humeri (UWBM 1641, 1643, 1659),
T5/100E, Unit 29, 1; T5/100E, Unit 28, IIB; Unit
20, RB. Three ulnae (UWBM 1240, 1241, 1647),
Unit 1 1, I-8 (2 bones); T5/100E, Unit 30,1 H.
Carpometacarpus (UWBM 1259), Unit 7, hA.
Manus digit (UWBM 1260), Unit 7, HA. Pelvis
(UWBM 1674), T9/500E, Unit 21, IIB. Three
tibiotarsi (UWBM 1252, 1253, 1652), Unit 6, hA (2
bones); T9/500E, Unit 23, IRA.
REMARKS. As reported by Steadman (1990),
these bones are larger than those of any species of
shearwater that resides today in tropical Polynesia.
They agree in all osteological details with bones of
Puffinus griseus, a species that probably migrates
through the Samoan region today (Harrison
1983:260, 420, Pratt et al. 1987:55), although there
are no records from American Samoa. The Sooty
Shearwater nests today only on islands off New
Zealand, southern Australia, and extreme southern
South America (Harrison 1983:260, 420). Most of
the nesting localities of P. griseus are temperate or
subantarctic, although in Australia nesting occurs at
least as far north as 320 40' S on the subtropical
Broughton Islands.
Three possible explanations for the unexpected
presence of Paufinus griseus on Ofu were proposed
by Steadman (1990): (1) the bones represent
migrant birds taken at sea by fishermen; (2) P.
griseus once resided on Ofu and, like other
procellarids, was extirpated through predation in
their nesting burrows by humans and rats; (3) the
bones represent an extinct, resident shearwater that
differs specifically or subspecifically from modem
P. griseiss but is osteologically very similar. Regard-
ing the first explanation, I am aware of no ethno-
graphic accounts of Polynesians capturing seabirds
while fishing. The data now available, particularly
that fifteen bones (nine different skeletal elements)
are osteologically indistinguishable from modem
skeletons of P. griseus, suggest that the second
explanation may be corect, even though the oceano-
graphic conditions near Ofu are wanmer than those
of the wannest localities where P. griseus nests
today. That this large shearwater (or an osteologi-
cally indistinguishable subspecies [explanation 3 is
not necessarily incompatible with explanation 2])
was once resident rather than migrant on Ofu is
indicated by the presence at To'aga of two bones of
P. griseus (UWBM 1645,1653) that, based upon
porosity of the external surface, are from volant
juveniles unlikely to have dispersed far from their
place of birth. The fonner residency of P. griseus on
Ofu is supported furtxr by the abun e of its
bones at To'aga (15 of 74, or 20% of identifiable
bones). Sooty Shearwaters are represented as well
among the few bird bones from a Lapita site on
*Niuatoputapu (Tonga), southwest of American
Samoa (Steadman 1990).
Pterodroma rostrata (Tahiti Petrel)
MATERIAL. Two mandibles (UWBM 1250,
1682), Unit 4, RB; T9/500E, Unit 21, IIB. Scapula
(UWBM 1684), 19/500E, Unit 21, IIB. Humenrs
(UWBM 1657), T9/500E, Unit 20, IIB. Ulna
(UWBM 1242), Unit 5, HA. Tarsometatarsus
(UWBM 1247), Unit 4, IIB.
REMARKS. This large petrel nests today at a
few widely scattered localities in te Marquesas
Islands, Tahiti, New Caledonia, and American
Samoa. Its presence on the latter is based upon a
colony of about 500 individuals discovered in 1976
in the montarn cloud forest of Ta'u, a single nesting
bird discovered on Tutuila in 1986, and a few birds
heard on Olosega in 1986. An onshore record in
1972 from Taveuni (Clunie et al. 1978) suggests th
Pterodroma rostrata may nest in Fiji as well.
Elsewhere in Polynesia, P. rostrata has been idend-
fied from archaeological sites on *Ua Huka and
Tahuata (Marquesas), *Huahine (Society Islands),
*Aitutaki (Cook Islands), *Lifuka and *Eua
(Tonga), Tikopia (Solomon Islands), and New
Caledonia.
Pterodroma sp. (unknown petrel,
size of P. heraldica)
MATERIAL Radius (UWBM 1670), T9/500E,
Unit 21, RB. Femur (UWBM 1676), Unit 21, RIB.
REMARKS. These specimens, although not
adequate for species-level identification, represent a
species of Pterodroma in the approximate size rae
Prehistoric Loss of Seabirds and Megapodes  221
of P. heraldica (Herald Petrel, which often is
considered conspecific with P. arminjonana) or
perhaps the slightly larger P. exlerna (Juan
Fernandez Petrel), both of which are significantly
smaller than P. rostrata. In American Samoa, P.
eterna is known ordy from uncommon sightings at
sea, while P. heraldica nests on TaVu. Regardless of
species-level identification, UWBM 1670 and 1676
indicate that a species of petrel other than P. rostrata
probably once nested on Ofu.
Procelladiidae sp. (unknown petrel/shearwater)
MATERIAL. Coracoid (UWBM 1633), T5/
100E, Unit 16,11. Five ulnae (UWBM 1243,1258,
1634, 1662, 1669), Unit 9, IIB; Unit 5, IIB; T5/
100E, Unit 16, I; T9/500E, Unit 20, IIIB; T9/500E,
Unit 21, fIB. Radius (UWBM 1249), Unit 4, IIB.
Two carpometacarpi (UWBM 1239, 1677), Unit 9,
UB; T9/500E, Unit 21, IIB.
REMARKS. These fragmentary bones cannot
be identified beyond the family level. They do not
represent a taxon separate from those already
identified.
Order Pelecaniformes
Family Sulidae
Sula sula (Red-footed Booby)
MATERIAL. Radius (UWBM 1656), T9/500E,
Unit 20, IIIC.
REMARKS. Sula sula nested on Ofu as
recently as 1975-76 (only twenty-five birds), but the
small colony no longer existed in 1986. This booby
still nested in 1986 on Tutuila and Rose islands. It is
widespread in tropical oceans. Other Polynesian
archaeological records of S. sula are from
*Henderson (Pitcaim Group); *Ua Huka, *Hiva Oa,
and *Tahuata (Marquesas); *Huahine (Society
Islands); *Aitutaki (Cook Islands); Niuatoputapu
(Tonga); and *Tikopia and *Anuta (Solomon
Islands).
Family Fregatidae
Fregata sp. (unknown frigatebird)
MATERIAL. Humerus (UWBM 1254), Unit
14, lIa-4. Ulna (UWBM 1675), T9/500E, Unit 21,
BB.
REMARKS. These specimens cannot be
distinguished from those of Fregata minor or F.
ariel, both of which visit but do not nest on Ofu
today (total of thirteen birds counted in 1975-76; still
recorded as a visitor in 1986). The bones of F.
minor males and F. ariel females are similar in size
and difficult to distinguish from each other
(Steadman et al. 1990). Both species of Fregata
occur fairly commonly in Polynesian archaeological
sites.
Order Charadriifonues
Family Laridae
Anous stolidus (Brown Noddy)
MATERIAL. Humerus (UWBM 1668),T9/
500E, TP 21, IIB.
REMARKS. This tern is common and wide-
spread today in American Samoa as well as most of
Polynesia. The Ofu population was ca. 500 in 1975-
76 and was not accurately estimated in 1986. Other
Polynesian archaeological records of Anous stolidus
are from Henderson (Pitcairn Group); Ua Pou, Ua
Huka, and Tahuata (Marquesas); Huahine (Society
Islands); Mangaia and Aitutaki (Cook Islands);
Niuatoputapu, Lifuka, and 'Eua (Tonga); 'Upolu
(Western Samoa); Tikopia and Anuta (Solomon
Islands); and Mussau (Papua New Guinea).
Gygis alba (Common Fairy-Tern)
MATERIAL. Humerus (UWBM 1631), T5/
100E, Unit 16,I.
REMARKS. This distinctive tern is common
and widespread today in most of Polynesia including
American Samoa. The Ofu population was ca. 100
in 1975-76, and at least 500 in 1986. Other
Polynesian archaeological records of Gygis alba are
from Henderson Island (Pitcairn Group), Ua Huka
and Tahuata (Marquesas), Huahine (Society Islands),
Mangaia (Cook Islands), and Niuatoputapu and 'Eua
(Tonga).
Sterninae sp. (unknown tern)
MATERIAL. Ulna (UWBM 1257), Unit 14,
llb-5.
REMARKS. This eroded, fragmentary speci-
men represents a tern smaller than either of the
above species. UWBM 1257 is approximately te
222     The To'aga Site
size of the ulna in Sterna swnatrana orAnous
minus, both of which occur today in American
Samoa, although only the latter nests on Ofu (just ten
birds in 1975-76, and perhaps about the same in
1986).
Order Clconiifonmes
Family Ardeidae
Egretta sacra (Pacific Reef-Heron)
MATERIAL. Coracoid (UWBM 1655), T9/
500E, Unit 22, 1.
REMARKS. Egreta sacra resides throughout
American Samoa and most of Polynesia today. It is
uncommon on Ofu, where only dark-phase birds
have been recorded. Other archaeological records
are from Nuku Hiva and Ua Huka (Marquesas),
Huahine (Society Islands), Mangaia and Aitutaki
(Cook Islands), and 'Eua (Tonga).
Order Charadriiformes
Family Scolopacidae
Nwnenius tahidensis (Bristle-thighed Curlew)
MATERIAL. Corcoid (UWBM 1635, 1636;
originally believed to be two separate bones), T5/
100E, Unit 15, IIID.
REMARKS. This widespread but rather
uncommon migrant shorebird has not been recorded
previously from Ofu, although undoubtedly it still
occurs there occasionally. American Samoan
records of Nwnenius tahlitensis are confined to
Tutuila, Ta'u, Swains, Rose, and Olosega. Other
Polynesian archaeological records for N. tahidensis
are from Henderson Island (Pitcaim Group), Ua
Huka (Marquesas), Huahine (Society Islands),
Mangaia and Aitutaki (Cook Islands), and Tikopia
(Solomon Islands).
During the autumn wing molt, thirteen of
twenty-nine individuals of N. tahtidensis captured on
Laysan (Leeward Hawaiian Islands) were flightless
(Marks et al. 1990). This adaptation would seem to
be viable only in a predator-free setting. It may have
led to reductions in the distribution and abundance of
N. ahitiensis following the human colonization of
Polynesia.
Order Galliformes
Family Megapodiidae
Megapodius sp. (Unknown Megapode)
MATERIAL. Ulna (UWBM 1637), T5/100E,
Unit 15, HID. Femur (UWBM 1649), Unit 30, HID.
REMARKS. These two fragmentary specimens
(fig. 14.2) are most similar quantitatively and
qualitatively to modem specimens of Megapodius
freycinet, a widespread species that now reaches its
eastern limit in Vanuatu. UWBM 1649 is indistin-
guishable qualitatively from the femur in modem M.
freycinet, while UWBM 1637 differs from the ulna
of M. freycinet in having a slightly deeper sulcus
radialis and in lacking a diagonal trough on the
cranial surface of tuberculum carpale.
Like the archaeological specimens of
Megapodiusfreycinet from *Tikopia (Solomon
Islands; Steadman et al. 1990), the specimens of
Megapodius from To'aga are at the extreme small
end of the size range of M.freycinet (table 14.2).
Among living subspecies of M.freycinet, there is no
indication that individual body size decreases
eastward in Oceania. For example, the tarsus (in
skins) of the Vanuatu population is not smaller han
that from Australia (table 14.3). The bones from
To'aga are larger than in M. pritchardi (confined to
Niuafo'ou, Tonga) and M. laperouse (found orny in
Palau and Mariana Islands), but much smaller than
in two extinct species recently described from late
Holocene archaeological and paleontological sites-
M. molistructor of New Caledonia and Lifuka and
M. alimentum of Lifuka and 'Eua (Balouet and Olson
1989; Steadman 1989b, pers. obs.). Megapodius
pritchardi is the only species of megapode that
survives east of Vanuatu.
Both To'aga specimens of Megapodius are from
one of the site's deepest and oldest strata, Layer IIID
in units 15/29/30. This suggests that megapodes
may have been lost from Ofu not long after prehis-
toric colonization of the island. These bones provide
the first unequivocal, well-documented record of a
megapode from Samoa. There is, however, compli-
cated historical evidence that a megapode, described
as M. stain by Gray (1861), may have existed in tbe
mid-i 800s on Upolu or Savai'i (Western Samoa).
Gray (1861) also described M. burnabyi from
Prehistoric Loss of Seabirds and Megapodes  223
Comparison of megapode bones: ulna in ventral aspect
(A, B) and femur in lateral aspect (C, D). A, C.
Megapodius cf.freycinet, archaeological specimens
(UWBM 1637, 1649), To'aga site, Ofu, American Samoa.
B, D. M. freycinet, modem specimen, USNM 557015,
male, Halmahera, Northern Moluccas. Scale = 20 mm.
Ha'apai (Tonga). The true identities of M. stairi and
Mf. burnabyi are uncertain because only a single egg
-ver was collected of each, and these two eggs
cannot be assigned unequivocally to M. pritchardi,
M.freycinet, or any other species of megapode
(Steadman 1991). One or both of these eggs may
represent the same species as the bones from Ofu,
which also are at the very lower limit of the size
range of M. freycinet and larger than, or at the
uppermost size limit of, M. pritchardi.
The original, handwritten data slip with the
holotypical egg of Megapodius burnabyi in the
British Museum notes that this egg was "called the
chief s egg' as they are only allowed to eat them."
Such a chiefly tabu might suggest rarity of the bird
and a knowledge that overexploitation, which
probably was the cause of the rarity, would eventu-
ally lead to extinction. Alternatively, being called
the "chief's egg" could suggest that megapode eggs
were prestigious trade items brought to Ha'apai from
another island. An extensive exchange network
operated among Fiji, Tonga, and Samoa in late
prehistoric and early historic times (Kirch 1984:238-
42; 1988:257-60). The Samoan voyages recorded by
Stair (1895) included Fiji and Tonga as well as much
of East Polynesia. Bennett (1862:247) noted that the
nesting grounds of M. pritchardi on Niuafo'ou were
"under the protection of the king or chief, and by his
permission only can the birds or eggs be procured."
Even if megapodes of the Fiji/Tonga/Samoa region
were confined by the nineteenth century to
Niuafo'ou, they would have been known to Tongans
in Ha'apai, as well as to Samoans.
Unless additional evidence comes forth, neither
C%
A     B
Figure 14.2
224     The To'aga Site
Table 14.2
Measurements (in mm) of the Femur and Ulna of Megapodius, with Mean,
Range, and Sample Size
Specimen                         Locality                      A        B        C       D        E
Megapodious sp.
UWBM 1637 (U)
Megapodius sp.
UWBM 1649 (U)
M. f cf. eremita
Lab #: 66, 68 (U)
M. f freycinet
USNM 556995-557002
557016, 557017 (M)
M. f freycinet
USNM 557006-557008
557010-557013, 557019-557022 (F)
M. f gilberti
USNM 226175, 226176 (F, M)
M. f cumingi
BM(NHl) 1862.2.10.2 (U)
M. f nicobariensis
USNM 19686, 19700 (M)
M. pritchardi
tJSNM 319633, 319634 (IJ)
Samoa: Ofu: To'aga Site
Samoa: Ofu: To'aga Site
New Ireland
N. Moluccas: Halmahera
N. Moluccas: Halmahera
Sulawesi
Philippines: island
unknown
Nicobar Islands
Tonga: Niuafo'ou
M. laperouse
USNM unnumbered (U)
M. alimenturm
UWBM 2100 (IJ)
N. Mariana Islands:
Rota: Payapai Cave
Tonga: 'Eua:
'Anatu (Ground-Dove Cave)
3.4       2.5
1         1
-        6.0       4.4
1         1
F = Female, M = Male, U = Sex unknown
Column Headings:
A. Femur: width at deepest proximo-lateral muscle scar
B. Femur: depth at deepest proximo-lateral muscle scar
C. Ulna: minimum width of shaft
D. Ulna: minimum depth of shaft
E. Ulna: width of distal end
6.5
1
7.7
1
4.4
1
3.3
1
8.4
1
6.8
1
4.9
1
7.3
1
8.2
7.6-9.1
10
8.3
7.7-8.8
10
7.4
7.3-7.6
2
5.0
4.7-5.3
10
4.9
4.6-5.4
10
4.3
4.3
2
3.8
3.3-4.1
10
3.7
3.5-4.1
10
3.2
3.2-3.3
2
7.4
7.0-7.9
10
7.3
7.0-7.8
10
6.5
6.5
2
8.2
7.5-8.7
10
8.2
7.6-8.8
10
7.2
6.9-7.5
2
8.1
1
8.3
8.0-8.6
2
5.6
5.4-5.7
2
8.0
1
5.2      4.1
1        1
7.4
1
8.3
8.2-8.4
2
6.0
5.9-6.1
2
5.0
4.9-5.0
2
3.4
3.3-3.5
2
3.8
3.6-4.0
2
2.8
2.8-2.9
2
7.8
7.6-7.9
2
5.2
5.0-5.3
2
5.4
1
Prehistoric Loss of Seabirds and Megapodes  225
Table 14.3
Tarsal Length (in mm) from Skins of Selected
Subspecies of Megapodius freycinet, with Mean,
Range, and Sample Size. Based upon Specimens
from BM(NH).
Subspecies
Tarsal Length
M. f layardi
Vanuatu: Santo, Vate
(F, 5M)
M. f eremita
Solomon Is.: San Cristobal,
Guadacanal, Bouganville
(3F, 2M, tI
M. f eremita
Papua New Guinea: New Britain
(3F, 2M, 3U)
M. f yorki
Australia: Queensland
(2F, 4M, 4tJ)
73.8 (71-76), 6
72.2 (67-77), 6
67.1 (62-71), 8
71.6 (70-74), 10
F = female, M = male, U = sex unknown
Megapodius stairi orM. burnabyi should be re-
garded as certain records of indigenous populations
of megapodes in nineteenth century Samoa or
Ha'apai, although this cannot be ruled out. The
survival of M. pritchardi on Niuafo'ou has been due
to chiefly control of exploiting eggs and birds at the
conspicuous nest mounds, as first described by
Bennett (1862). The people of Niuafo'ou must have
realized that conserving megapodes, which probably
occurred nowhere else in the region, would help to
maintain their share of commerce in the Samoa-
Tonga-Fiji trade network.
Family Phasianidae
Gallus gallus (Chicken)
MATERIAL. Sternum (UWBM 1261), Unit 11,
1. Coracoid (UWBM 1654), T9/500E, Unit 23, IIIB.
Scapula (UWBM 1663), T9/500E, Unit 20, HIB.
Ulna (UWBM 1255), Unit 14, Illa-4. Radius
(UJWBM 1660), T9/500E, Unit 20, IhB. Two pelves
(UWBM 1665, 1686, 1688; the last two originally
believed to be separate bones), T9/500E, Unit 20,
RIB; T9/500E, Unit 21, IIB. Two femora (UWBM
1666, 1667), T9/500E, Unit 20, hIB. Four tibiotarsi
(IJWBM 1632, 1644, 1658, 1661), T5/100E, Unit
16, 1; T5/1OOE, Unit 29, IIIB; T9/500E, Unit 20, IIB;
T9/500E, Unit 20, IhB. Two tarsometatarsi
(UWBM 1646, 1683), Unit 29, rIB; Unit 21, II.
REMARKS. Feral and/or domestic populations
of Gallus gallus occur neatly throughout Polynesia,
including all inhabited Samoan islands. All chickens
recorded on Ofu in 1986 were near human habitation
and not from deep within forests. This non-native
species has been found thugh virtually all of
Polynesia in archaeological sites of any age, except
that it is absent from all sites on Henderson Island
(Schubel and Steadman 1989). Chicken bones occur
throughout the To'aga sequence, but are especially
well represented in the Layer IIIB occupation in
Units 20/23 along Transect 9: This indicates that G.
gallus was a commonly eaten bird during the
Ancestral Polynesian phase, ca. 2500 yr B.P.
226     The To'aga Site
Order Columbiformes
Family Columbidae
Gallicolwnba stairii (Shy Ground-Dove)
MATERIAL. Humerus (UWBM 1638), T5/
100E, Unit 15, 111D. Ulna (UWBM 1658), Unit 30,
IIID. Tarsometatarsus (UWBM 1690), T5/100E,
Unit 29, IIID.
REMARKS. This species occurs only in very
old deposits at To'aga, primarily Layer IIID in Units
29/30. Gallicolwnba stairii is extremely rare on Ofu
today, with a very roughly estimated 100 birds
surviving in 1975-76. Only two or three ground-
doves were seen on Ofu during the 1986 surveys; no
population estimate was made. Within American
Samoa, only perhaps on Olosega does another small
population of G. stairii survive. Similar declines or
losses of populations of G. stairii have occurred in
Tonga, where the only other archaeological record of
G. stairii is from *'Eua.
DISCUSSION
Virtually all of the bird bones from To'aga are
broken, often with both articular ends missing. Most
of these breaks are not fresh, although often it is
difficult to distinguish whether human or sedimen-
tary processes have caused the breakage. A few
bones are rounded, suggesting some post-mortem
sedimentary transport. The breakage and rounding
might indicate that the calcareous sands at To'aga
represent a somewhat higher energy deposit than the
calcareous sands at certain other Polynesian archaeo-
logical sites, such as Hane (Ua Huka, Marquesas) or
Tongoleleka (Lifuka, Tonga). Among the shearwa-
ters and petrels, however, a systematic butchering
technique is suggested by the fairly consistent
pattern of both ends of the humerus, ulna, and
tibiotarsus being broken off. Two of the chicken
bones had been chewed by rats. None of the bird
bones seems to have been modified into recogniz-
able artifacts.
Among indigenous, resident species recorded
from the To'aga site, five of ten seabirds and one of
three landbirds are extirpated on Ofu (table 14.1). At
least two of the surviving species (Sterninae sp.,
Gallicolwnba stairit) exist today on Ofu only in very
small, threatened populations. Should these species
be lost from Ofu, the proportion of bones of extir-
pated species at To'aga would increase from 85%
(table 14.1) to 93%.
The majority of bird bones from To'aga (46 of
74, or 62%) are of at least five species of petrels or
shearwaters (table 14.1), none of which nests on OfN
today (Amerson et al. 1982a:90). Only two of hse
species (Audubon's Shearwater and Tahiti Petrel)
are known certainly to nest today anywhere in
American Samoa. As seems to be case throughout
Polynesia (Steadman 1989a, Dye and Steadman
1990), the island of Ofu had a diverse and probably
abundant seabird fauna when humans first arrived.
In the case of Ofu, not a single species of procellariid
has survived the three millennia of human occupa-
tion.
Compared to avian assemblages from other
Polynesian archaeological sites, the dominance of
procellariids at To'aga would characterize a fairly
early site, i.e., one that dates to within the first
thousand years of human occupation (Dye and
Steadman 1990). When compared to sites that seem
to represent the initial human occupation of an
island, however, such as the Hane site (Ua Huka,
Marquesas; Steadman 1989a) orTongoleleka site
(Lifuka, Tonga; Steadman 1989b), the To'aga site's
lower percentage and diversity of bones from native
landbirds and higher percentage of chicken bones
would suggest that this site may not represent the
first 500 years of human occupation of Ofu. This
suggestion is compatible with the radiocarbon
chronology at To'aga which indicates occupation of
the site from about 2800 to 1900 yr B.P. (see Kirch,
chapter 6).
Although the bird bones from To'aga provide
much new data on the prehistoric distribution of
Samoan birds, a bone sample about an order of
magnitude larger would be necessary to provide a
fairly complete picture of the past birdlife of Ofu.
That six of the fifteen taxa of birds from To'aga are
represented by only a single bone indicates that more
species await discovery if a larger bone sample were
available. The 1987 sample of twenty-three bones
yielded six taxa, which increased to fifteen taxa with
the addition of fifty-one bones from the 1989
excavations. The point of diminishing returns is
difficult to determine, however, from comparison
with assemblages from other sites. For example,
thirty-five species of birds are represented in a
sample of ca. 350 bones from the Fa'ahia site
(Huahine, Society Islands; Steadman and Pahlavan
Prehistoric Loss of Seabirds and Megapodes  227
1992), while thirty-nine species of birds ae repre-
sented in a sample of ca. 12,000 bones from the
Hane site (Ua Huka, Marquesas) (Steadman 1989a,
pers. obs.). Twelve of the species from Hane are
known from either one or two bones. Another gauge
of the incompleteness of the To'aga avifauna is the
low percentage of species in the modem avifauna of
Ofu tha are represented at the site. In this case, only
two of sixteen possible species of resident landbirds
(12.5%) ar represented.
The To'aga site has given us an intriguing
introduction to Samoa's prehistoric birdlife. Our
understanding of the relationship between native
birds and the first human inhabitants of Samoa
undoubtedly will improve as more early archaeologi-
cal sites are discovered and carefully excavated.
ACKNOWLEDGEMENTS
This research was supported in part by National
Science Foundation grant BSR-8607535. I than T.
L. Hunt and P. V. Kirch for inviting me to study the
bird bones from To'aga. L. Nagaoka and D. S.
Pahlavan assisted in curation. S. L. Olson and M. P.
Walters kindly shared information on megapodes.
For comments on the manuscript, I thank T. L. Hunt,
P. V. Kirch, N. G. Miller, S. L. Olson, and M. P.
Walters. This is contribution number 670 of the
New York State Museum and Science Service.
REFERENCES CITED
Amerson, A. B., Jr., W. A. Whistler, and T. D.
Schwaner 1982a. Wildlife and Wildlife Habitat
of American Samoa. 1. Environment and
Ecology. U.S. Dept. of Interior, Fish and
Wildlife Service, Washington D.C.
. 1982b. Wild!ife and Wild0lfe Habitat of Ameri-
can Samoa. H. Accounts of Flora and Fauna.
U. S. Dept. of Interior, Fish and Wildlife
Service, Washington D.C.
Balouet, J. C., and S. L. Olson 1989. Fossil birds
from late Quatemary deposits in New
Caledonia. Smithsonian Contributions to
Zoology 469.
Bennett, G. 1862. [Letter from Dr. G. Bennett on
Didunculus, Tallegalla, and Megapodius.]
Proceedings of the Zoological Society of
London 1862:24648.
Clunie, F., F. C. Kinsky, and J. A. F. Jenkins 1978.
New bird records from the Fiji Archipelago.
Notornis 25:118-27.
Dye, T., and D. W. Steadman 1990. Polynesian
ancestors and their animal world. American
Scientist 78:207-215.
Engbring, J., and F. L. Ramsey 1989. A 1986 Survey
of the Forest Birds of American Samoa. U.S.
Dept. of Interior, Fish and Wildlife Service,
Washington D.C.
Gray, G. R 1861. [pub. April 1862] List of species
composing the family Megapodiidae, with
descriptions of new species, and some account
of the habits of the species. Proceedings of the
Zoological Society of London 1861:288-96.
Hanison, P. 1983. Seabirds: An Identification Guide.
Houghton Mifflin Co., Boston.
Hunt, T. L., and P. V. Kirch 1988. An archaeological
survey of the Manu'a Islands, American
Samoa. Journal of the Polynesian Society
97:153-83.
Kirch, P. V. 1984. The Evolution of the Polynesian
Chiefdoms. Cambridge: Cambridge University
Press.
1988. Niuatoputapu: The Prehistory of a
Polynesian Chiefdom. Thomas Burke Memo-
rial Washington State Museum Monograph No.
5.
Kirch, P. V., T. L. Hunt, and J. Tyler 1989. A
radiocarbon sequence from the To'aga site, Ofu
Island, American Samoa. Radiocarbon 31:7-
13.
Kirch, P. V., T. L. Hunt, L. Nagaoka, and J. Tyler
1990. An ancestral Polynesian occupation site
at To'aga, Ofu Island, American Samoa.
Archaeology in Oceania 25:1-15.
Marks, J. S., R. L. Redmond, P. Hendricks, R. B.
Clapp, and R. E. Gill, Jr. 1990. Notes on
longevity and Sightlessness in Bristle-thighed
Curlews. Auk 107:779-81.
Pratt, H. D., P. L. Bruner, and D. G. Berrett 1987. A
Field Guide to the Birds of Hawaii and the
Tropical Pacific. Princeton: Princeton Univer-
sity Press.
Schubel, S. E., and D. W. Steadman 1989. More bird
bones from Polynesian archaeological sites on
Henderson Island, Pitcaim Group, South
Pacific. Atoll Research Bulletin 325:1-18.
Stair, J. B. 1895. Flotsam and jetsam from the great
Ocean; or, summary of early Samoan voyages
and settlements. Journal of the Polynesian
228     The Tolaga Site
Society 4:99-131.
Steadman, D. W. 1989a. Extinction of birds in
Eastern Polynesia: A review of the record, and
comparisons with other Pacific island groups.
Journal ofArchaeological Science 16:177-205.
1989b. New species and records of birds (Aves:
Megapodiidae, Columbidae) from an archeo-
logical site on Lifuka, Tonga. Proceedings of
the Biological Society of Washington 102:537-
52.
. 1990. Archaeological bird bones from Ofu
[Manu'a, American Samoa]: Extirpation of
shearwaters and petrels. Archaeology in
Oceania 25:14-15.
1991. The identity and taxonomic status of
Megapodius stairi and M. burnabyi (Aves:
Megapodiidae). Proceedings of the Biological
Society of Washington 104:870-77.
Steadman, D. W., and D. S. Pahlavan 1992. Prehis-
toric exploitation and extinction of birds on
Huahine, Society Islands, French Polynesia.
Geoarchaeology 7:449-83.
Steadman, D. W., D. S. Pahlavan, and P. V. Kirch
1990. Extinction, biogeography, and human
exploitation of birds on Tikopia and Anuta,
Polynesian outliers in the Solomon Islands.
Occasional Papers of the Bishop Museum
30:118-53. Honolulu: Bemice P. Bishop
Museum.